postgresql/contrib/btree_gist/btree_gist--1.2.sql

1571 lines
40 KiB
MySQL
Raw Normal View History

/* contrib/btree_gist/btree_gist--1.2.sql */
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "CREATE EXTENSION btree_gist" to load this file. \quit
CREATE FUNCTION gbtreekey4_in(cstring)
RETURNS gbtreekey4
AS 'MODULE_PATHNAME', 'gbtreekey_in'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbtreekey4_out(gbtreekey4)
RETURNS cstring
AS 'MODULE_PATHNAME', 'gbtreekey_out'
LANGUAGE C IMMUTABLE STRICT;
CREATE TYPE gbtreekey4 (
INTERNALLENGTH = 4,
INPUT = gbtreekey4_in,
OUTPUT = gbtreekey4_out
);
CREATE FUNCTION gbtreekey8_in(cstring)
RETURNS gbtreekey8
AS 'MODULE_PATHNAME', 'gbtreekey_in'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbtreekey8_out(gbtreekey8)
RETURNS cstring
AS 'MODULE_PATHNAME', 'gbtreekey_out'
LANGUAGE C IMMUTABLE STRICT;
CREATE TYPE gbtreekey8 (
INTERNALLENGTH = 8,
INPUT = gbtreekey8_in,
OUTPUT = gbtreekey8_out
);
CREATE FUNCTION gbtreekey16_in(cstring)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME', 'gbtreekey_in'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbtreekey16_out(gbtreekey16)
RETURNS cstring
AS 'MODULE_PATHNAME', 'gbtreekey_out'
LANGUAGE C IMMUTABLE STRICT;
CREATE TYPE gbtreekey16 (
INTERNALLENGTH = 16,
INPUT = gbtreekey16_in,
OUTPUT = gbtreekey16_out
);
CREATE FUNCTION gbtreekey32_in(cstring)
RETURNS gbtreekey32
AS 'MODULE_PATHNAME', 'gbtreekey_in'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbtreekey32_out(gbtreekey32)
RETURNS cstring
AS 'MODULE_PATHNAME', 'gbtreekey_out'
LANGUAGE C IMMUTABLE STRICT;
CREATE TYPE gbtreekey32 (
INTERNALLENGTH = 32,
INPUT = gbtreekey32_in,
OUTPUT = gbtreekey32_out
);
CREATE FUNCTION gbtreekey_var_in(cstring)
RETURNS gbtreekey_var
AS 'MODULE_PATHNAME', 'gbtreekey_in'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbtreekey_var_out(gbtreekey_var)
RETURNS cstring
AS 'MODULE_PATHNAME', 'gbtreekey_out'
LANGUAGE C IMMUTABLE STRICT;
CREATE TYPE gbtreekey_var (
INTERNALLENGTH = VARIABLE,
INPUT = gbtreekey_var_in,
OUTPUT = gbtreekey_var_out,
STORAGE = EXTENDED
);
--distance operators
CREATE FUNCTION cash_dist(money, money)
RETURNS money
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = money,
RIGHTARG = money,
PROCEDURE = cash_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION date_dist(date, date)
RETURNS int4
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = date,
RIGHTARG = date,
PROCEDURE = date_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION float4_dist(float4, float4)
RETURNS float4
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = float4,
RIGHTARG = float4,
PROCEDURE = float4_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION float8_dist(float8, float8)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = float8,
RIGHTARG = float8,
PROCEDURE = float8_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION int2_dist(int2, int2)
RETURNS int2
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = int2,
RIGHTARG = int2,
PROCEDURE = int2_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION int4_dist(int4, int4)
RETURNS int4
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = int4,
RIGHTARG = int4,
PROCEDURE = int4_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION int8_dist(int8, int8)
RETURNS int8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = int8,
RIGHTARG = int8,
PROCEDURE = int8_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION interval_dist(interval, interval)
RETURNS interval
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = interval,
RIGHTARG = interval,
PROCEDURE = interval_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION oid_dist(oid, oid)
RETURNS oid
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = oid,
RIGHTARG = oid,
PROCEDURE = oid_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION time_dist(time, time)
RETURNS interval
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = time,
RIGHTARG = time,
PROCEDURE = time_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION ts_dist(timestamp, timestamp)
RETURNS interval
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = timestamp,
RIGHTARG = timestamp,
PROCEDURE = ts_dist,
COMMUTATOR = '<->'
);
CREATE FUNCTION tstz_dist(timestamptz, timestamptz)
RETURNS interval
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE OPERATOR <-> (
LEFTARG = timestamptz,
RIGHTARG = timestamptz,
PROCEDURE = tstz_dist,
COMMUTATOR = '<->'
);
--
--
--
-- oid ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_oid_consistent(internal,oid,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_oid_distance(internal,oid,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_oid_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_oid_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_decompress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_var_decompress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_var_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_oid_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_oid_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_oid_union(internal, internal)
RETURNS gbtreekey8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_oid_same(gbtreekey8, gbtreekey8, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_oid_ops
DEFAULT FOR TYPE oid USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_oid_consistent (internal, oid, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_oid_union (internal, internal),
FUNCTION 3 gbt_oid_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_oid_penalty (internal, internal, internal),
FUNCTION 6 gbt_oid_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_oid_same (gbtreekey8, gbtreekey8, internal),
STORAGE gbtreekey8;
-- Add operators that are new in 9.1. We do it like this, leaving them
-- "loose" in the operator family rather than bound into the opclass, because
-- that's the only state that can be reproduced during an upgrade from 9.0.
ALTER OPERATOR FAMILY gist_oid_ops USING gist ADD
OPERATOR 6 <> (oid, oid) ,
OPERATOR 15 <-> (oid, oid) FOR ORDER BY pg_catalog.oid_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (oid, oid) gbt_oid_distance (internal, oid, int2, oid, internal) ,
-- Also add support function for index-only-scans, added in 9.5.
FUNCTION 9 (oid, oid) gbt_oid_fetch (internal) ;
--
--
--
-- int2 ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_int2_consistent(internal,int2,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int2_distance(internal,int2,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int2_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int2_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int2_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int2_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int2_union(internal, internal)
RETURNS gbtreekey4
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int2_same(gbtreekey4, gbtreekey4, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_int2_ops
DEFAULT FOR TYPE int2 USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_int2_consistent (internal, int2, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_int2_union (internal, internal),
FUNCTION 3 gbt_int2_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_int2_penalty (internal, internal, internal),
FUNCTION 6 gbt_int2_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_int2_same (gbtreekey4, gbtreekey4, internal),
STORAGE gbtreekey4;
ALTER OPERATOR FAMILY gist_int2_ops USING gist ADD
OPERATOR 6 <> (int2, int2) ,
OPERATOR 15 <-> (int2, int2) FOR ORDER BY pg_catalog.integer_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (int2, int2) gbt_int2_distance (internal, int2, int2, oid, internal) ,
FUNCTION 9 (int2, int2) gbt_int2_fetch (internal) ;
--
--
--
-- int4 ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_int4_consistent(internal,int4,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int4_distance(internal,int4,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int4_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int4_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int4_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int4_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int4_union(internal, internal)
RETURNS gbtreekey8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int4_same(gbtreekey8, gbtreekey8, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_int4_ops
DEFAULT FOR TYPE int4 USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_int4_consistent (internal, int4, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_int4_union (internal, internal),
FUNCTION 3 gbt_int4_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_int4_penalty (internal, internal, internal),
FUNCTION 6 gbt_int4_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_int4_same (gbtreekey8, gbtreekey8, internal),
STORAGE gbtreekey8;
ALTER OPERATOR FAMILY gist_int4_ops USING gist ADD
OPERATOR 6 <> (int4, int4) ,
OPERATOR 15 <-> (int4, int4) FOR ORDER BY pg_catalog.integer_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (int4, int4) gbt_int4_distance (internal, int4, int2, oid, internal) ,
FUNCTION 9 (int4, int4) gbt_int4_fetch (internal) ;
--
--
--
-- int8 ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_int8_consistent(internal,int8,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int8_distance(internal,int8,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int8_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int8_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int8_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_int8_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int8_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_int8_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_int8_ops
DEFAULT FOR TYPE int8 USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_int8_consistent (internal, int8, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_int8_union (internal, internal),
FUNCTION 3 gbt_int8_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_int8_penalty (internal, internal, internal),
FUNCTION 6 gbt_int8_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_int8_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_int8_ops USING gist ADD
OPERATOR 6 <> (int8, int8) ,
OPERATOR 15 <-> (int8, int8) FOR ORDER BY pg_catalog.integer_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (int8, int8) gbt_int8_distance (internal, int8, int2, oid, internal) ,
FUNCTION 9 (int8, int8) gbt_int8_fetch (internal) ;
--
--
--
-- float4 ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_float4_consistent(internal,float4,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float4_distance(internal,float4,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float4_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float4_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float4_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float4_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float4_union(internal, internal)
RETURNS gbtreekey8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float4_same(gbtreekey8, gbtreekey8, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_float4_ops
DEFAULT FOR TYPE float4 USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_float4_consistent (internal, float4, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_float4_union (internal, internal),
FUNCTION 3 gbt_float4_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_float4_penalty (internal, internal, internal),
FUNCTION 6 gbt_float4_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_float4_same (gbtreekey8, gbtreekey8, internal),
STORAGE gbtreekey8;
ALTER OPERATOR FAMILY gist_float4_ops USING gist ADD
OPERATOR 6 <> (float4, float4) ,
OPERATOR 15 <-> (float4, float4) FOR ORDER BY pg_catalog.float_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (float4, float4) gbt_float4_distance (internal, float4, int2, oid, internal) ,
FUNCTION 9 (float4, float4) gbt_float4_fetch (internal) ;
--
--
--
-- float8 ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_float8_consistent(internal,float8,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float8_distance(internal,float8,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float8_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float8_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float8_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_float8_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float8_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_float8_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_float8_ops
DEFAULT FOR TYPE float8 USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_float8_consistent (internal, float8, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_float8_union (internal, internal),
FUNCTION 3 gbt_float8_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_float8_penalty (internal, internal, internal),
FUNCTION 6 gbt_float8_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_float8_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_float8_ops USING gist ADD
OPERATOR 6 <> (float8, float8) ,
OPERATOR 15 <-> (float8, float8) FOR ORDER BY pg_catalog.float_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (float8, float8) gbt_float8_distance (internal, float8, int2, oid, internal) ,
FUNCTION 9 (float8, float8) gbt_float8_fetch (internal) ;
--
--
--
-- timestamp ops
--
--
--
CREATE FUNCTION gbt_ts_consistent(internal,timestamp,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_ts_distance(internal,timestamp,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_tstz_consistent(internal,timestamptz,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_tstz_distance(internal,timestamptz,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_ts_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_tstz_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_ts_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_ts_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_ts_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_ts_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_ts_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_timestamp_ops
DEFAULT FOR TYPE timestamp USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_ts_consistent (internal, timestamp, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_ts_union (internal, internal),
FUNCTION 3 gbt_ts_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_ts_penalty (internal, internal, internal),
FUNCTION 6 gbt_ts_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_ts_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_timestamp_ops USING gist ADD
OPERATOR 6 <> (timestamp, timestamp) ,
OPERATOR 15 <-> (timestamp, timestamp) FOR ORDER BY pg_catalog.interval_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (timestamp, timestamp) gbt_ts_distance (internal, timestamp, int2, oid, internal) ,
FUNCTION 9 (timestamp, timestamp) gbt_ts_fetch (internal) ;
-- Create the operator class
CREATE OPERATOR CLASS gist_timestamptz_ops
DEFAULT FOR TYPE timestamptz USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_tstz_consistent (internal, timestamptz, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_ts_union (internal, internal),
FUNCTION 3 gbt_tstz_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_ts_penalty (internal, internal, internal),
FUNCTION 6 gbt_ts_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_ts_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_timestamptz_ops USING gist ADD
OPERATOR 6 <> (timestamptz, timestamptz) ,
OPERATOR 15 <-> (timestamptz, timestamptz) FOR ORDER BY pg_catalog.interval_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (timestamptz, timestamptz) gbt_tstz_distance (internal, timestamptz, int2, oid, internal) ,
FUNCTION 9 (timestamptz, timestamptz) gbt_ts_fetch (internal) ;
--
--
--
-- time ops
--
--
--
CREATE FUNCTION gbt_time_consistent(internal,time,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_time_distance(internal,time,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_timetz_consistent(internal,timetz,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_time_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_timetz_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_time_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_time_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_time_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_time_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_time_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_time_ops
DEFAULT FOR TYPE time USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_time_consistent (internal, time, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_time_union (internal, internal),
FUNCTION 3 gbt_time_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_time_penalty (internal, internal, internal),
FUNCTION 6 gbt_time_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_time_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_time_ops USING gist ADD
OPERATOR 6 <> (time, time) ,
OPERATOR 15 <-> (time, time) FOR ORDER BY pg_catalog.interval_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (time, time) gbt_time_distance (internal, time, int2, oid, internal) ,
FUNCTION 9 (time, time) gbt_time_fetch (internal) ;
CREATE OPERATOR CLASS gist_timetz_ops
DEFAULT FOR TYPE timetz USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_timetz_consistent (internal, timetz, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_time_union (internal, internal),
FUNCTION 3 gbt_timetz_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_time_penalty (internal, internal, internal),
FUNCTION 6 gbt_time_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_time_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_timetz_ops USING gist ADD
OPERATOR 6 <> (timetz, timetz) ;
-- no 'fetch' function, as the compress function is lossy.
--
--
--
-- date ops
--
--
--
CREATE FUNCTION gbt_date_consistent(internal,date,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_date_distance(internal,date,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_date_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_date_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_date_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_date_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_date_union(internal, internal)
RETURNS gbtreekey8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_date_same(gbtreekey8, gbtreekey8, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_date_ops
DEFAULT FOR TYPE date USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_date_consistent (internal, date, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_date_union (internal, internal),
FUNCTION 3 gbt_date_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_date_penalty (internal, internal, internal),
FUNCTION 6 gbt_date_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_date_same (gbtreekey8, gbtreekey8, internal),
STORAGE gbtreekey8;
ALTER OPERATOR FAMILY gist_date_ops USING gist ADD
OPERATOR 6 <> (date, date) ,
OPERATOR 15 <-> (date, date) FOR ORDER BY pg_catalog.integer_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (date, date) gbt_date_distance (internal, date, int2, oid, internal) ,
FUNCTION 9 (date, date) gbt_date_fetch (internal) ;
--
--
--
-- interval ops
--
--
--
CREATE FUNCTION gbt_intv_consistent(internal,interval,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_intv_distance(internal,interval,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_intv_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_intv_decompress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_intv_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_intv_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_intv_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_intv_union(internal, internal)
RETURNS gbtreekey32
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_intv_same(gbtreekey32, gbtreekey32, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_interval_ops
DEFAULT FOR TYPE interval USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_intv_consistent (internal, interval, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_intv_union (internal, internal),
FUNCTION 3 gbt_intv_compress (internal),
FUNCTION 4 gbt_intv_decompress (internal),
FUNCTION 5 gbt_intv_penalty (internal, internal, internal),
FUNCTION 6 gbt_intv_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_intv_same (gbtreekey32, gbtreekey32, internal),
STORAGE gbtreekey32;
ALTER OPERATOR FAMILY gist_interval_ops USING gist ADD
OPERATOR 6 <> (interval, interval) ,
OPERATOR 15 <-> (interval, interval) FOR ORDER BY pg_catalog.interval_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (interval, interval) gbt_intv_distance (internal, interval, int2, oid, internal) ,
FUNCTION 9 (interval, interval) gbt_intv_fetch (internal) ;
--
--
--
-- cash ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_cash_consistent(internal,money,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_cash_distance(internal,money,int2,oid,internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_cash_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_cash_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_cash_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_cash_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_cash_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_cash_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_cash_ops
DEFAULT FOR TYPE money USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_cash_consistent (internal, money, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_cash_union (internal, internal),
FUNCTION 3 gbt_cash_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_cash_penalty (internal, internal, internal),
FUNCTION 6 gbt_cash_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_cash_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_cash_ops USING gist ADD
OPERATOR 6 <> (money, money) ,
OPERATOR 15 <-> (money, money) FOR ORDER BY pg_catalog.money_ops ,
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 8 (money, money) gbt_cash_distance (internal, money, int2, oid, internal) ,
FUNCTION 9 (money, money) gbt_cash_fetch (internal) ;
--
--
--
-- macaddr ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_macad_consistent(internal,macaddr,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_macad_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_macad_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_macad_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_macad_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_macad_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_macad_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_macaddr_ops
DEFAULT FOR TYPE macaddr USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_macad_consistent (internal, macaddr, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_macad_union (internal, internal),
FUNCTION 3 gbt_macad_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_macad_penalty (internal, internal, internal),
FUNCTION 6 gbt_macad_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_macad_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_macaddr_ops USING gist ADD
OPERATOR 6 <> (macaddr, macaddr) ,
FUNCTION 9 (macaddr, macaddr) gbt_macad_fetch (internal);
--
--
--
-- text/ bpchar ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_text_consistent(internal,text,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bpchar_consistent(internal,bpchar,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_text_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bpchar_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_text_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_text_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_text_union(internal, internal)
RETURNS gbtreekey_var
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_text_same(gbtreekey_var, gbtreekey_var, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_text_ops
DEFAULT FOR TYPE text USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_text_consistent (internal, text, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_text_union (internal, internal),
FUNCTION 3 gbt_text_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_text_penalty (internal, internal, internal),
FUNCTION 6 gbt_text_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_text_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_text_ops USING gist ADD
OPERATOR 6 <> (text, text) ,
FUNCTION 9 (text, text) gbt_var_fetch (internal) ;
---- Create the operator class
CREATE OPERATOR CLASS gist_bpchar_ops
DEFAULT FOR TYPE bpchar USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_bpchar_consistent (internal, bpchar , int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_text_union (internal, internal),
FUNCTION 3 gbt_bpchar_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_text_penalty (internal, internal, internal),
FUNCTION 6 gbt_text_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_text_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_bpchar_ops USING gist ADD
OPERATOR 6 <> (bpchar, bpchar) ,
FUNCTION 9 (bpchar, bpchar) gbt_var_fetch (internal) ;
--
--
-- bytea ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_bytea_consistent(internal,bytea,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bytea_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bytea_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bytea_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_bytea_union(internal, internal)
RETURNS gbtreekey_var
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_bytea_same(gbtreekey_var, gbtreekey_var, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_bytea_ops
DEFAULT FOR TYPE bytea USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_bytea_consistent (internal, bytea, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_bytea_union (internal, internal),
FUNCTION 3 gbt_bytea_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_bytea_penalty (internal, internal, internal),
FUNCTION 6 gbt_bytea_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_bytea_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_bytea_ops USING gist ADD
OPERATOR 6 <> (bytea, bytea) ,
FUNCTION 9 (bytea, bytea) gbt_var_fetch (internal) ;
--
--
--
-- numeric ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_numeric_consistent(internal,numeric,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_numeric_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_numeric_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_numeric_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_numeric_union(internal, internal)
RETURNS gbtreekey_var
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_numeric_same(gbtreekey_var, gbtreekey_var, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_numeric_ops
DEFAULT FOR TYPE numeric USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_numeric_consistent (internal, numeric, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_numeric_union (internal, internal),
FUNCTION 3 gbt_numeric_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_numeric_penalty (internal, internal, internal),
FUNCTION 6 gbt_numeric_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_numeric_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_numeric_ops USING gist ADD
OPERATOR 6 <> (numeric, numeric) ,
FUNCTION 9 (numeric, numeric) gbt_var_fetch (internal) ;
--
--
-- bit ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_bit_consistent(internal,bit,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bit_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bit_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_bit_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_bit_union(internal, internal)
RETURNS gbtreekey_var
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_bit_same(gbtreekey_var, gbtreekey_var, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_bit_ops
DEFAULT FOR TYPE bit USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_bit_consistent (internal, bit, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_bit_union (internal, internal),
FUNCTION 3 gbt_bit_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_bit_penalty (internal, internal, internal),
FUNCTION 6 gbt_bit_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_bit_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_bit_ops USING gist ADD
OPERATOR 6 <> (bit, bit) ,
FUNCTION 9 (bit, bit) gbt_var_fetch (internal) ;
-- Create the operator class
CREATE OPERATOR CLASS gist_vbit_ops
DEFAULT FOR TYPE varbit USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_bit_consistent (internal, bit, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_bit_union (internal, internal),
FUNCTION 3 gbt_bit_compress (internal),
FUNCTION 4 gbt_var_decompress (internal),
FUNCTION 5 gbt_bit_penalty (internal, internal, internal),
FUNCTION 6 gbt_bit_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_bit_same (gbtreekey_var, gbtreekey_var, internal),
STORAGE gbtreekey_var;
ALTER OPERATOR FAMILY gist_vbit_ops USING gist ADD
OPERATOR 6 <> (varbit, varbit) ,
FUNCTION 9 (varbit, varbit) gbt_var_fetch (internal) ;
--
--
--
-- inet/cidr ops
--
--
--
-- define the GiST support methods
CREATE FUNCTION gbt_inet_consistent(internal,inet,int2,oid,internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_inet_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_inet_penalty(internal,internal,internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
CREATE FUNCTION gbt_inet_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_inet_union(internal, internal)
RETURNS gbtreekey16
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
CREATE FUNCTION gbt_inet_same(gbtreekey16, gbtreekey16, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C IMMUTABLE STRICT;
-- Create the operator class
CREATE OPERATOR CLASS gist_inet_ops
DEFAULT FOR TYPE inet USING gist
AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 gbt_inet_consistent (internal, inet, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_inet_union (internal, internal),
FUNCTION 3 gbt_inet_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_inet_penalty (internal, internal, internal),
FUNCTION 6 gbt_inet_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_inet_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_inet_ops USING gist ADD
OPERATOR 6 <> (inet, inet) ;
-- no fetch support, the compress function is lossy
-- Create the operator class
CREATE OPERATOR CLASS gist_cidr_ops
DEFAULT FOR TYPE cidr USING gist
AS
OPERATOR 1 < (inet, inet) ,
OPERATOR 2 <= (inet, inet) ,
OPERATOR 3 = (inet, inet) ,
OPERATOR 4 >= (inet, inet) ,
OPERATOR 5 > (inet, inet) ,
FUNCTION 1 gbt_inet_consistent (internal, inet, int2, oid, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 2 gbt_inet_union (internal, internal),
FUNCTION 3 gbt_inet_compress (internal),
FUNCTION 4 gbt_decompress (internal),
FUNCTION 5 gbt_inet_penalty (internal, internal, internal),
FUNCTION 6 gbt_inet_picksplit (internal, internal),
Fix assorted inconsistencies in GiST opclass support function declarations. The conventions specified by the GiST SGML documentation were widely ignored. For example, the strategy-number argument for "consistent" and "distance" functions is specified to be a smallint, but most of the built-in support functions declared it as an integer, and for that matter the core code passed it using Int32GetDatum not Int16GetDatum. None of that makes any real difference at runtime, but it's quite confusing for newcomers to the code, and it makes it very hard to write an amvalidate() function that checks support function signatures. So let's try to instill some consistency here. Another similar issue is that the "query" argument is not of a single well-defined type, but could have different types depending on the strategy (corresponding to search operators with different righthand-side argument types). Some of the functions threw up their hands and declared the query argument as being of "internal" type, which surely isn't right ("any" would have been more appropriate); but the majority position seemed to be to declare it as being of the indexed data type, corresponding to a search operator with both input types the same. So I've specified a convention that that's what to do always. Also, the result of the "union" support function actually must be of the index's storage type, but the documentation suggested declaring it to return "internal", and some of the functions followed that. Standardize on telling the truth, instead. Similarly, standardize on declaring the "same" function's inputs as being of the storage type, not "internal". Also, somebody had forgotten to add the "recheck" argument to both the documentation of the "distance" support function and all of their SQL declarations, even though the C code was happily using that argument. Clean that up too. Fix up some other omissions in the docs too, such as documenting that union's second input argument is vestigial. So far as the errors in core function declarations go, we can just fix pg_proc.h and bump catversion. Adjusting the erroneous declarations in contrib modules is more debatable: in principle any change in those scripts should involve an extension version bump, which is a pain. However, since these changes are purely cosmetic and make no functional difference, I think we can get away without doing that.
2016-01-20 01:04:32 +08:00
FUNCTION 7 gbt_inet_same (gbtreekey16, gbtreekey16, internal),
STORAGE gbtreekey16;
ALTER OPERATOR FAMILY gist_cidr_ops USING gist ADD
OPERATOR 6 <> (inet, inet) ;
-- no fetch support, the compress function is lossy