mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2024-12-03 04:12:10 +08:00
492d29ea1c
This patch splits the TRY_CATCH macro into three, so that we go from this: ~~~ volatile gdb_exception ex; TRY_CATCH (ex, RETURN_MASK_ERROR) { } if (ex.reason < 0) { } ~~~ to this: ~~~ TRY { } CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH ~~~ Thus, we'll be getting rid of the local volatile exception object, and declaring the caught exception in the catch block. This allows reimplementing TRY/CATCH in terms of C++ exceptions when building in C++ mode, while still allowing to build GDB in C mode (using setjmp/longjmp), as a transition step. TBC, after this patch, is it _not_ valid to have code between the TRY and the CATCH blocks, like: TRY { } // some code here. CATCH (ex, RETURN_MASK_ERROR) { } END_CATCH Just like it isn't valid to do that with C++'s native try/catch. By switching to creating the exception object inside the CATCH block scope, we can get rid of all the explicitly allocated volatile exception objects all over the tree, and map the CATCH block more directly to C++'s catch blocks. The majority of the TRY_CATCH -> TRY+CATCH+END_CATCH conversion was done with a script, rerun from scratch at every rebase, no manual editing involved. After the mechanical conversion, a few places needed manual intervention, to fix preexisting cases where we were using the exception object outside of the TRY_CATCH block, and cases where we were using "else" after a 'if (ex.reason) < 0)' [a CATCH after this patch]. The result was folded into this patch so that GDB still builds at each incremental step. END_CATCH is necessary for two reasons: First, because we name the exception object in the CATCH block, which requires creating a scope, which in turn must be closed somewhere. Declaring the exception variable in the initializer field of a for block, like: #define CATCH(EXCEPTION, mask) \ for (struct gdb_exception EXCEPTION; \ exceptions_state_mc_catch (&EXCEPTION, MASK); \ EXCEPTION = exception_none) would avoid needing END_CATCH, but alas, in C mode, we build with C90, which doesn't allow mixed declarations and code. Second, because when TRY/CATCH are wired to real C++ try/catch, as long as we need to handle cleanup chains, even if there's no CATCH block that wants to catch the exception, we need for stop at every frame in the unwind chain and run cleanups, then rethrow. That will be done in END_CATCH. After we require C++, we'll still need TRY/CATCH/END_CATCH until cleanups are completely phased out -- TRY/CATCH in C++ mode will save/restore the current cleanup chain, like in C mode, and END_CATCH catches otherwise uncaugh exceptions, runs cleanups and rethrows, so that C++ cleanups and exceptions can coexist. IMO, this still makes the TRY/CATCH code look a bit more like a newcomer would expect, so IMO worth it even if we weren't considering C++. gdb/ChangeLog. 2015-03-07 Pedro Alves <palves@redhat.com> * common/common-exceptions.c (struct catcher) <exception>: No longer a pointer to volatile exception. Now an exception value. <mask>: Delete field. (exceptions_state_mc_init): Remove all parameters. Adjust. (exceptions_state_mc): No longer pop the catcher here. (exceptions_state_mc_catch): New function. (throw_exception): Adjust. * common/common-exceptions.h (exceptions_state_mc_init): Remove all parameters. (exceptions_state_mc_catch): Declare. (TRY_CATCH): Rename to ... (TRY): ... this. Remove EXCEPTION and MASK parameters. (CATCH, END_CATCH): New. All callers adjusted. gdb/gdbserver/ChangeLog: 2015-03-07 Pedro Alves <palves@redhat.com> Adjust all callers of TRY_CATCH to use TRY/CATCH/END_CATCH instead.
947 lines
26 KiB
C
947 lines
26 KiB
C
/* Support for printing Ada types for GDB, the GNU debugger.
|
||
Copyright (C) 1986-2015 Free Software Foundation, Inc.
|
||
|
||
This file is part of GDB.
|
||
|
||
This program is free software; you can redistribute it and/or modify
|
||
it under the terms of the GNU General Public License as published by
|
||
the Free Software Foundation; either version 3 of the License, or
|
||
(at your option) any later version.
|
||
|
||
This program is distributed in the hope that it will be useful,
|
||
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
GNU General Public License for more details.
|
||
|
||
You should have received a copy of the GNU General Public License
|
||
along with this program. If not, see <http://www.gnu.org/licenses/>. */
|
||
|
||
#include "defs.h"
|
||
#include "gdb_obstack.h"
|
||
#include "bfd.h" /* Binary File Description */
|
||
#include "symtab.h"
|
||
#include "gdbtypes.h"
|
||
#include "expression.h"
|
||
#include "value.h"
|
||
#include "gdbcore.h"
|
||
#include "target.h"
|
||
#include "command.h"
|
||
#include "gdbcmd.h"
|
||
#include "language.h"
|
||
#include "demangle.h"
|
||
#include "c-lang.h"
|
||
#include "typeprint.h"
|
||
#include "ada-lang.h"
|
||
#include <ctype.h>
|
||
|
||
static int print_selected_record_field_types (struct type *, struct type *,
|
||
int, int,
|
||
struct ui_file *, int, int,
|
||
const struct type_print_options *);
|
||
|
||
static int print_record_field_types (struct type *, struct type *,
|
||
struct ui_file *, int, int,
|
||
const struct type_print_options *);
|
||
|
||
|
||
|
||
static char *name_buffer;
|
||
static int name_buffer_len;
|
||
|
||
/* The (decoded) Ada name of TYPE. This value persists until the
|
||
next call. */
|
||
|
||
static char *
|
||
decoded_type_name (struct type *type)
|
||
{
|
||
if (ada_type_name (type) == NULL)
|
||
return NULL;
|
||
else
|
||
{
|
||
const char *raw_name = ada_type_name (type);
|
||
char *s, *q;
|
||
|
||
if (name_buffer == NULL || name_buffer_len <= strlen (raw_name))
|
||
{
|
||
name_buffer_len = 16 + 2 * strlen (raw_name);
|
||
name_buffer = xrealloc (name_buffer, name_buffer_len);
|
||
}
|
||
strcpy (name_buffer, raw_name);
|
||
|
||
s = (char *) strstr (name_buffer, "___");
|
||
if (s != NULL)
|
||
*s = '\0';
|
||
|
||
s = name_buffer + strlen (name_buffer) - 1;
|
||
while (s > name_buffer && (s[0] != '_' || s[-1] != '_'))
|
||
s -= 1;
|
||
|
||
if (s == name_buffer)
|
||
return name_buffer;
|
||
|
||
if (!islower (s[1]))
|
||
return NULL;
|
||
|
||
for (s = q = name_buffer; *s != '\0'; q += 1)
|
||
{
|
||
if (s[0] == '_' && s[1] == '_')
|
||
{
|
||
*q = '.';
|
||
s += 2;
|
||
}
|
||
else
|
||
{
|
||
*q = *s;
|
||
s += 1;
|
||
}
|
||
}
|
||
*q = '\0';
|
||
return name_buffer;
|
||
}
|
||
}
|
||
|
||
/* Return nonzero if TYPE is a subrange type, and its bounds
|
||
are identical to the bounds of its subtype. */
|
||
|
||
static int
|
||
type_is_full_subrange_of_target_type (struct type *type)
|
||
{
|
||
struct type *subtype;
|
||
|
||
if (TYPE_CODE (type) != TYPE_CODE_RANGE)
|
||
return 0;
|
||
|
||
subtype = TYPE_TARGET_TYPE (type);
|
||
if (subtype == NULL)
|
||
return 0;
|
||
|
||
if (is_dynamic_type (type))
|
||
return 0;
|
||
|
||
if (ada_discrete_type_low_bound (type)
|
||
!= ada_discrete_type_low_bound (subtype))
|
||
return 0;
|
||
|
||
if (ada_discrete_type_high_bound (type)
|
||
!= ada_discrete_type_high_bound (subtype))
|
||
return 0;
|
||
|
||
return 1;
|
||
}
|
||
|
||
/* Print TYPE on STREAM, preferably as a range if BOUNDS_PREFERED_P
|
||
is nonzero. */
|
||
|
||
static void
|
||
print_range (struct type *type, struct ui_file *stream,
|
||
int bounds_prefered_p)
|
||
{
|
||
if (!bounds_prefered_p)
|
||
{
|
||
/* Try stripping all TYPE_CODE_RANGE layers whose bounds
|
||
are identical to the bounds of their subtype. When
|
||
the bounds of both types match, it can allow us to
|
||
print a range using the name of its base type, which
|
||
is easier to read. For instance, we would print...
|
||
|
||
array (character) of ...
|
||
|
||
... instead of...
|
||
|
||
array ('["00"]' .. '["ff"]') of ... */
|
||
while (type_is_full_subrange_of_target_type (type))
|
||
type = TYPE_TARGET_TYPE (type);
|
||
}
|
||
|
||
switch (TYPE_CODE (type))
|
||
{
|
||
case TYPE_CODE_RANGE:
|
||
case TYPE_CODE_ENUM:
|
||
{
|
||
struct type *target_type;
|
||
LONGEST lo = 0, hi = 0; /* init for gcc -Wall */
|
||
int got_error = 0;
|
||
|
||
target_type = TYPE_TARGET_TYPE (type);
|
||
if (target_type == NULL)
|
||
target_type = type;
|
||
|
||
TRY
|
||
{
|
||
lo = ada_discrete_type_low_bound (type);
|
||
hi = ada_discrete_type_high_bound (type);
|
||
}
|
||
CATCH (e, RETURN_MASK_ERROR)
|
||
{
|
||
/* This can happen when the range is dynamic. Sometimes,
|
||
resolving dynamic property values requires us to have
|
||
access to an actual object, which is not available
|
||
when the user is using the "ptype" command on a type.
|
||
Print the range as an unbounded range. */
|
||
fprintf_filtered (stream, "<>");
|
||
got_error = 1;
|
||
}
|
||
END_CATCH
|
||
|
||
if (!got_error)
|
||
{
|
||
ada_print_scalar (target_type, lo, stream);
|
||
fprintf_filtered (stream, " .. ");
|
||
ada_print_scalar (target_type, hi, stream);
|
||
}
|
||
}
|
||
break;
|
||
default:
|
||
fprintf_filtered (stream, "%.*s",
|
||
ada_name_prefix_len (TYPE_NAME (type)),
|
||
TYPE_NAME (type));
|
||
break;
|
||
}
|
||
}
|
||
|
||
/* Print the number or discriminant bound at BOUNDS+*N on STREAM, and
|
||
set *N past the bound and its delimiter, if any. */
|
||
|
||
static void
|
||
print_range_bound (struct type *type, char *bounds, int *n,
|
||
struct ui_file *stream)
|
||
{
|
||
LONGEST B;
|
||
|
||
if (ada_scan_number (bounds, *n, &B, n))
|
||
{
|
||
/* STABS decodes all range types which bounds are 0 .. -1 as
|
||
unsigned integers (ie. the type code is TYPE_CODE_INT, not
|
||
TYPE_CODE_RANGE). Unfortunately, ada_print_scalar() relies
|
||
on the unsigned flag to determine whether the bound should
|
||
be printed as a signed or an unsigned value. This causes
|
||
the upper bound of the 0 .. -1 range types to be printed as
|
||
a very large unsigned number instead of -1.
|
||
To workaround this stabs deficiency, we replace the TYPE by NULL
|
||
to indicate default output when we detect that the bound is negative,
|
||
and the type is a TYPE_CODE_INT. The bound is negative when
|
||
'm' is the last character of the number scanned in BOUNDS. */
|
||
if (bounds[*n - 1] == 'm' && TYPE_CODE (type) == TYPE_CODE_INT)
|
||
type = NULL;
|
||
ada_print_scalar (type, B, stream);
|
||
if (bounds[*n] == '_')
|
||
*n += 2;
|
||
}
|
||
else
|
||
{
|
||
int bound_len;
|
||
char *bound = bounds + *n;
|
||
char *pend;
|
||
|
||
pend = strstr (bound, "__");
|
||
if (pend == NULL)
|
||
*n += bound_len = strlen (bound);
|
||
else
|
||
{
|
||
bound_len = pend - bound;
|
||
*n += bound_len + 2;
|
||
}
|
||
fprintf_filtered (stream, "%.*s", bound_len, bound);
|
||
}
|
||
}
|
||
|
||
/* Assuming NAME[0 .. NAME_LEN-1] is the name of a range type, print
|
||
the value (if found) of the bound indicated by SUFFIX ("___L" or
|
||
"___U") according to the ___XD conventions. */
|
||
|
||
static void
|
||
print_dynamic_range_bound (struct type *type, const char *name, int name_len,
|
||
const char *suffix, struct ui_file *stream)
|
||
{
|
||
static char *name_buf = NULL;
|
||
static size_t name_buf_len = 0;
|
||
LONGEST B;
|
||
int OK;
|
||
|
||
GROW_VECT (name_buf, name_buf_len, name_len + strlen (suffix) + 1);
|
||
strncpy (name_buf, name, name_len);
|
||
strcpy (name_buf + name_len, suffix);
|
||
|
||
B = get_int_var_value (name_buf, &OK);
|
||
if (OK)
|
||
ada_print_scalar (type, B, stream);
|
||
else
|
||
fprintf_filtered (stream, "?");
|
||
}
|
||
|
||
/* Print RAW_TYPE as a range type, using any bound information
|
||
following the GNAT encoding (if available).
|
||
|
||
If BOUNDS_PREFERED_P is nonzero, force the printing of the range
|
||
using its bounds. Otherwise, try printing the range without
|
||
printing the value of the bounds, if possible (this is only
|
||
considered a hint, not a guaranty). */
|
||
|
||
static void
|
||
print_range_type (struct type *raw_type, struct ui_file *stream,
|
||
int bounds_prefered_p)
|
||
{
|
||
const char *name;
|
||
struct type *base_type;
|
||
const char *subtype_info;
|
||
|
||
gdb_assert (raw_type != NULL);
|
||
name = TYPE_NAME (raw_type);
|
||
gdb_assert (name != NULL);
|
||
|
||
if (TYPE_CODE (raw_type) == TYPE_CODE_RANGE)
|
||
base_type = TYPE_TARGET_TYPE (raw_type);
|
||
else
|
||
base_type = raw_type;
|
||
|
||
subtype_info = strstr (name, "___XD");
|
||
if (subtype_info == NULL)
|
||
print_range (raw_type, stream, bounds_prefered_p);
|
||
else
|
||
{
|
||
int prefix_len = subtype_info - name;
|
||
char *bounds_str;
|
||
int n;
|
||
|
||
subtype_info += 5;
|
||
bounds_str = strchr (subtype_info, '_');
|
||
n = 1;
|
||
|
||
if (*subtype_info == 'L')
|
||
{
|
||
print_range_bound (base_type, bounds_str, &n, stream);
|
||
subtype_info += 1;
|
||
}
|
||
else
|
||
print_dynamic_range_bound (base_type, name, prefix_len, "___L",
|
||
stream);
|
||
|
||
fprintf_filtered (stream, " .. ");
|
||
|
||
if (*subtype_info == 'U')
|
||
print_range_bound (base_type, bounds_str, &n, stream);
|
||
else
|
||
print_dynamic_range_bound (base_type, name, prefix_len, "___U",
|
||
stream);
|
||
}
|
||
}
|
||
|
||
/* Print enumerated type TYPE on STREAM. */
|
||
|
||
static void
|
||
print_enum_type (struct type *type, struct ui_file *stream)
|
||
{
|
||
int len = TYPE_NFIELDS (type);
|
||
int i;
|
||
LONGEST lastval;
|
||
|
||
fprintf_filtered (stream, "(");
|
||
wrap_here (" ");
|
||
|
||
lastval = 0;
|
||
for (i = 0; i < len; i++)
|
||
{
|
||
QUIT;
|
||
if (i)
|
||
fprintf_filtered (stream, ", ");
|
||
wrap_here (" ");
|
||
fputs_filtered (ada_enum_name (TYPE_FIELD_NAME (type, i)), stream);
|
||
if (lastval != TYPE_FIELD_ENUMVAL (type, i))
|
||
{
|
||
fprintf_filtered (stream, " => %s",
|
||
plongest (TYPE_FIELD_ENUMVAL (type, i)));
|
||
lastval = TYPE_FIELD_ENUMVAL (type, i);
|
||
}
|
||
lastval += 1;
|
||
}
|
||
fprintf_filtered (stream, ")");
|
||
}
|
||
|
||
/* Print representation of Ada fixed-point type TYPE on STREAM. */
|
||
|
||
static void
|
||
print_fixed_point_type (struct type *type, struct ui_file *stream)
|
||
{
|
||
DOUBLEST delta = ada_delta (type);
|
||
DOUBLEST small = ada_fixed_to_float (type, 1.0);
|
||
|
||
if (delta < 0.0)
|
||
fprintf_filtered (stream, "delta ??");
|
||
else
|
||
{
|
||
fprintf_filtered (stream, "delta %g", (double) delta);
|
||
if (delta != small)
|
||
fprintf_filtered (stream, " <'small = %g>", (double) small);
|
||
}
|
||
}
|
||
|
||
/* Print simple (constrained) array type TYPE on STREAM. LEVEL is the
|
||
recursion (indentation) level, in case the element type itself has
|
||
nested structure, and SHOW is the number of levels of internal
|
||
structure to show (see ada_print_type). */
|
||
|
||
static void
|
||
print_array_type (struct type *type, struct ui_file *stream, int show,
|
||
int level, const struct type_print_options *flags)
|
||
{
|
||
int bitsize;
|
||
int n_indices;
|
||
|
||
if (ada_is_constrained_packed_array_type (type))
|
||
type = ada_coerce_to_simple_array_type (type);
|
||
|
||
bitsize = 0;
|
||
fprintf_filtered (stream, "array (");
|
||
|
||
if (type == NULL)
|
||
{
|
||
fprintf_filtered (stream, _("<undecipherable array type>"));
|
||
return;
|
||
}
|
||
|
||
n_indices = -1;
|
||
if (ada_is_simple_array_type (type))
|
||
{
|
||
struct type *range_desc_type;
|
||
struct type *arr_type;
|
||
|
||
range_desc_type = ada_find_parallel_type (type, "___XA");
|
||
ada_fixup_array_indexes_type (range_desc_type);
|
||
|
||
bitsize = 0;
|
||
if (range_desc_type == NULL)
|
||
{
|
||
for (arr_type = type; TYPE_CODE (arr_type) == TYPE_CODE_ARRAY;
|
||
arr_type = TYPE_TARGET_TYPE (arr_type))
|
||
{
|
||
if (arr_type != type)
|
||
fprintf_filtered (stream, ", ");
|
||
print_range (TYPE_INDEX_TYPE (arr_type), stream,
|
||
0 /* bounds_prefered_p */);
|
||
if (TYPE_FIELD_BITSIZE (arr_type, 0) > 0)
|
||
bitsize = TYPE_FIELD_BITSIZE (arr_type, 0);
|
||
}
|
||
}
|
||
else
|
||
{
|
||
int k;
|
||
|
||
n_indices = TYPE_NFIELDS (range_desc_type);
|
||
for (k = 0, arr_type = type;
|
||
k < n_indices;
|
||
k += 1, arr_type = TYPE_TARGET_TYPE (arr_type))
|
||
{
|
||
if (k > 0)
|
||
fprintf_filtered (stream, ", ");
|
||
print_range_type (TYPE_FIELD_TYPE (range_desc_type, k),
|
||
stream, 0 /* bounds_prefered_p */);
|
||
if (TYPE_FIELD_BITSIZE (arr_type, 0) > 0)
|
||
bitsize = TYPE_FIELD_BITSIZE (arr_type, 0);
|
||
}
|
||
}
|
||
}
|
||
else
|
||
{
|
||
int i, i0;
|
||
|
||
for (i = i0 = ada_array_arity (type); i > 0; i -= 1)
|
||
fprintf_filtered (stream, "%s<>", i == i0 ? "" : ", ");
|
||
}
|
||
|
||
fprintf_filtered (stream, ") of ");
|
||
wrap_here ("");
|
||
ada_print_type (ada_array_element_type (type, n_indices), "", stream,
|
||
show == 0 ? 0 : show - 1, level + 1, flags);
|
||
if (bitsize > 0)
|
||
fprintf_filtered (stream, " <packed: %d-bit elements>", bitsize);
|
||
}
|
||
|
||
/* Print the choices encoded by field FIELD_NUM of variant-part TYPE on
|
||
STREAM, assuming that VAL_TYPE (if non-NULL) is the type of the
|
||
values. Return non-zero if the field is an encoding of
|
||
discriminant values, as in a standard variant record, and 0 if the
|
||
field is not so encoded (as happens with single-component variants
|
||
in types annotated with pragma Unchecked_Variant). */
|
||
|
||
static int
|
||
print_choices (struct type *type, int field_num, struct ui_file *stream,
|
||
struct type *val_type)
|
||
{
|
||
int have_output;
|
||
int p;
|
||
const char *name = TYPE_FIELD_NAME (type, field_num);
|
||
|
||
have_output = 0;
|
||
|
||
/* Skip over leading 'V': NOTE soon to be obsolete. */
|
||
if (name[0] == 'V')
|
||
{
|
||
if (!ada_scan_number (name, 1, NULL, &p))
|
||
goto Huh;
|
||
}
|
||
else
|
||
p = 0;
|
||
|
||
while (1)
|
||
{
|
||
switch (name[p])
|
||
{
|
||
default:
|
||
goto Huh;
|
||
case '_':
|
||
case '\0':
|
||
fprintf_filtered (stream, " =>");
|
||
return 1;
|
||
case 'S':
|
||
case 'R':
|
||
case 'O':
|
||
if (have_output)
|
||
fprintf_filtered (stream, " | ");
|
||
have_output = 1;
|
||
break;
|
||
}
|
||
|
||
switch (name[p])
|
||
{
|
||
case 'S':
|
||
{
|
||
LONGEST W;
|
||
|
||
if (!ada_scan_number (name, p + 1, &W, &p))
|
||
goto Huh;
|
||
ada_print_scalar (val_type, W, stream);
|
||
break;
|
||
}
|
||
case 'R':
|
||
{
|
||
LONGEST L, U;
|
||
|
||
if (!ada_scan_number (name, p + 1, &L, &p)
|
||
|| name[p] != 'T' || !ada_scan_number (name, p + 1, &U, &p))
|
||
goto Huh;
|
||
ada_print_scalar (val_type, L, stream);
|
||
fprintf_filtered (stream, " .. ");
|
||
ada_print_scalar (val_type, U, stream);
|
||
break;
|
||
}
|
||
case 'O':
|
||
fprintf_filtered (stream, "others");
|
||
p += 1;
|
||
break;
|
||
}
|
||
}
|
||
|
||
Huh:
|
||
fprintf_filtered (stream, "?? =>");
|
||
return 0;
|
||
}
|
||
|
||
/* Assuming that field FIELD_NUM of TYPE represents variants whose
|
||
discriminant is contained in OUTER_TYPE, print its components on STREAM.
|
||
LEVEL is the recursion (indentation) level, in case any of the fields
|
||
themselves have nested structure, and SHOW is the number of levels of
|
||
internal structure to show (see ada_print_type). For this purpose,
|
||
fields nested in a variant part are taken to be at the same level as
|
||
the fields immediately outside the variant part. */
|
||
|
||
static void
|
||
print_variant_clauses (struct type *type, int field_num,
|
||
struct type *outer_type, struct ui_file *stream,
|
||
int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
int i;
|
||
struct type *var_type, *par_type;
|
||
struct type *discr_type;
|
||
|
||
var_type = TYPE_FIELD_TYPE (type, field_num);
|
||
discr_type = ada_variant_discrim_type (var_type, outer_type);
|
||
|
||
if (TYPE_CODE (var_type) == TYPE_CODE_PTR)
|
||
{
|
||
var_type = TYPE_TARGET_TYPE (var_type);
|
||
if (var_type == NULL || TYPE_CODE (var_type) != TYPE_CODE_UNION)
|
||
return;
|
||
}
|
||
|
||
par_type = ada_find_parallel_type (var_type, "___XVU");
|
||
if (par_type != NULL)
|
||
var_type = par_type;
|
||
|
||
for (i = 0; i < TYPE_NFIELDS (var_type); i += 1)
|
||
{
|
||
fprintf_filtered (stream, "\n%*swhen ", level + 4, "");
|
||
if (print_choices (var_type, i, stream, discr_type))
|
||
{
|
||
if (print_record_field_types (TYPE_FIELD_TYPE (var_type, i),
|
||
outer_type, stream, show, level + 4,
|
||
flags)
|
||
<= 0)
|
||
fprintf_filtered (stream, " null;");
|
||
}
|
||
else
|
||
print_selected_record_field_types (var_type, outer_type, i, i,
|
||
stream, show, level + 4, flags);
|
||
}
|
||
}
|
||
|
||
/* Assuming that field FIELD_NUM of TYPE is a variant part whose
|
||
discriminants are contained in OUTER_TYPE, print a description of it
|
||
on STREAM. LEVEL is the recursion (indentation) level, in case any of
|
||
the fields themselves have nested structure, and SHOW is the number of
|
||
levels of internal structure to show (see ada_print_type). For this
|
||
purpose, fields nested in a variant part are taken to be at the same
|
||
level as the fields immediately outside the variant part. */
|
||
|
||
static void
|
||
print_variant_part (struct type *type, int field_num, struct type *outer_type,
|
||
struct ui_file *stream, int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
fprintf_filtered (stream, "\n%*scase %s is", level + 4, "",
|
||
ada_variant_discrim_name
|
||
(TYPE_FIELD_TYPE (type, field_num)));
|
||
print_variant_clauses (type, field_num, outer_type, stream, show,
|
||
level + 4, flags);
|
||
fprintf_filtered (stream, "\n%*send case;", level + 4, "");
|
||
}
|
||
|
||
/* Print a description on STREAM of the fields FLD0 through FLD1 in
|
||
record or union type TYPE, whose discriminants are in OUTER_TYPE.
|
||
LEVEL is the recursion (indentation) level, in case any of the
|
||
fields themselves have nested structure, and SHOW is the number of
|
||
levels of internal structure to show (see ada_print_type). Does
|
||
not print parent type information of TYPE. Returns 0 if no fields
|
||
printed, -1 for an incomplete type, else > 0. Prints each field
|
||
beginning on a new line, but does not put a new line at end. */
|
||
|
||
static int
|
||
print_selected_record_field_types (struct type *type, struct type *outer_type,
|
||
int fld0, int fld1,
|
||
struct ui_file *stream, int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
int i, flds;
|
||
|
||
flds = 0;
|
||
|
||
if (fld0 > fld1 && TYPE_STUB (type))
|
||
return -1;
|
||
|
||
for (i = fld0; i <= fld1; i += 1)
|
||
{
|
||
QUIT;
|
||
|
||
if (ada_is_parent_field (type, i) || ada_is_ignored_field (type, i))
|
||
;
|
||
else if (ada_is_wrapper_field (type, i))
|
||
flds += print_record_field_types (TYPE_FIELD_TYPE (type, i), type,
|
||
stream, show, level, flags);
|
||
else if (ada_is_variant_part (type, i))
|
||
{
|
||
print_variant_part (type, i, outer_type, stream, show, level, flags);
|
||
flds = 1;
|
||
}
|
||
else
|
||
{
|
||
flds += 1;
|
||
fprintf_filtered (stream, "\n%*s", level + 4, "");
|
||
ada_print_type (TYPE_FIELD_TYPE (type, i),
|
||
TYPE_FIELD_NAME (type, i),
|
||
stream, show - 1, level + 4, flags);
|
||
fprintf_filtered (stream, ";");
|
||
}
|
||
}
|
||
|
||
return flds;
|
||
}
|
||
|
||
/* Print a description on STREAM of all fields of record or union type
|
||
TYPE, as for print_selected_record_field_types, above. */
|
||
|
||
static int
|
||
print_record_field_types (struct type *type, struct type *outer_type,
|
||
struct ui_file *stream, int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
return print_selected_record_field_types (type, outer_type,
|
||
0, TYPE_NFIELDS (type) - 1,
|
||
stream, show, level, flags);
|
||
}
|
||
|
||
|
||
/* Print record type TYPE on STREAM. LEVEL is the recursion (indentation)
|
||
level, in case the element type itself has nested structure, and SHOW is
|
||
the number of levels of internal structure to show (see ada_print_type). */
|
||
|
||
static void
|
||
print_record_type (struct type *type0, struct ui_file *stream, int show,
|
||
int level, const struct type_print_options *flags)
|
||
{
|
||
struct type *parent_type;
|
||
struct type *type;
|
||
|
||
type = ada_find_parallel_type (type0, "___XVE");
|
||
if (type == NULL)
|
||
type = type0;
|
||
|
||
parent_type = ada_parent_type (type);
|
||
if (ada_type_name (parent_type) != NULL)
|
||
{
|
||
const char *parent_name = decoded_type_name (parent_type);
|
||
|
||
/* If we fail to decode the parent type name, then use the parent
|
||
type name as is. Not pretty, but should never happen except
|
||
when the debugging info is incomplete or incorrect. This
|
||
prevents a crash trying to print a NULL pointer. */
|
||
if (parent_name == NULL)
|
||
parent_name = ada_type_name (parent_type);
|
||
fprintf_filtered (stream, "new %s with record", parent_name);
|
||
}
|
||
else if (parent_type == NULL && ada_is_tagged_type (type, 0))
|
||
fprintf_filtered (stream, "tagged record");
|
||
else
|
||
fprintf_filtered (stream, "record");
|
||
|
||
if (show < 0)
|
||
fprintf_filtered (stream, " ... end record");
|
||
else
|
||
{
|
||
int flds;
|
||
|
||
flds = 0;
|
||
if (parent_type != NULL && ada_type_name (parent_type) == NULL)
|
||
flds += print_record_field_types (parent_type, parent_type,
|
||
stream, show, level, flags);
|
||
flds += print_record_field_types (type, type, stream, show, level,
|
||
flags);
|
||
|
||
if (flds > 0)
|
||
fprintf_filtered (stream, "\n%*send record", level, "");
|
||
else if (flds < 0)
|
||
fprintf_filtered (stream, _(" <incomplete type> end record"));
|
||
else
|
||
fprintf_filtered (stream, " null; end record");
|
||
}
|
||
}
|
||
|
||
/* Print the unchecked union type TYPE in something resembling Ada
|
||
format on STREAM. LEVEL is the recursion (indentation) level
|
||
in case the element type itself has nested structure, and SHOW is the
|
||
number of levels of internal structure to show (see ada_print_type). */
|
||
static void
|
||
print_unchecked_union_type (struct type *type, struct ui_file *stream,
|
||
int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
if (show < 0)
|
||
fprintf_filtered (stream, "record (?) is ... end record");
|
||
else if (TYPE_NFIELDS (type) == 0)
|
||
fprintf_filtered (stream, "record (?) is null; end record");
|
||
else
|
||
{
|
||
int i;
|
||
|
||
fprintf_filtered (stream, "record (?) is\n%*scase ? is", level + 4, "");
|
||
|
||
for (i = 0; i < TYPE_NFIELDS (type); i += 1)
|
||
{
|
||
fprintf_filtered (stream, "\n%*swhen ? =>\n%*s", level + 8, "",
|
||
level + 12, "");
|
||
ada_print_type (TYPE_FIELD_TYPE (type, i),
|
||
TYPE_FIELD_NAME (type, i),
|
||
stream, show - 1, level + 12, flags);
|
||
fprintf_filtered (stream, ";");
|
||
}
|
||
|
||
fprintf_filtered (stream, "\n%*send case;\n%*send record",
|
||
level + 4, "", level, "");
|
||
}
|
||
}
|
||
|
||
|
||
|
||
/* Print function or procedure type TYPE on STREAM. Make it a header
|
||
for function or procedure NAME if NAME is not null. */
|
||
|
||
static void
|
||
print_func_type (struct type *type, struct ui_file *stream, const char *name,
|
||
const struct type_print_options *flags)
|
||
{
|
||
int i, len = TYPE_NFIELDS (type);
|
||
|
||
if (TYPE_CODE (TYPE_TARGET_TYPE (type)) == TYPE_CODE_VOID)
|
||
fprintf_filtered (stream, "procedure");
|
||
else
|
||
fprintf_filtered (stream, "function");
|
||
|
||
if (name != NULL && name[0] != '\0')
|
||
fprintf_filtered (stream, " %s", name);
|
||
|
||
if (len > 0)
|
||
{
|
||
fprintf_filtered (stream, " (");
|
||
for (i = 0; i < len; i += 1)
|
||
{
|
||
if (i > 0)
|
||
{
|
||
fputs_filtered ("; ", stream);
|
||
wrap_here (" ");
|
||
}
|
||
fprintf_filtered (stream, "a%d: ", i + 1);
|
||
ada_print_type (TYPE_FIELD_TYPE (type, i), "", stream, -1, 0,
|
||
flags);
|
||
}
|
||
fprintf_filtered (stream, ")");
|
||
}
|
||
|
||
if (TYPE_CODE (TYPE_TARGET_TYPE (type)) != TYPE_CODE_VOID)
|
||
{
|
||
fprintf_filtered (stream, " return ");
|
||
ada_print_type (TYPE_TARGET_TYPE (type), "", stream, 0, 0, flags);
|
||
}
|
||
}
|
||
|
||
|
||
/* Print a description of a type TYPE0.
|
||
Output goes to STREAM (via stdio).
|
||
If VARSTRING is a non-empty string, print as an Ada variable/field
|
||
declaration.
|
||
SHOW+1 is the maximum number of levels of internal type structure
|
||
to show (this applies to record types, enumerated types, and
|
||
array types).
|
||
SHOW is the number of levels of internal type structure to show
|
||
when there is a type name for the SHOWth deepest level (0th is
|
||
outer level).
|
||
When SHOW<0, no inner structure is shown.
|
||
LEVEL indicates level of recursion (for nested definitions). */
|
||
|
||
void
|
||
ada_print_type (struct type *type0, const char *varstring,
|
||
struct ui_file *stream, int show, int level,
|
||
const struct type_print_options *flags)
|
||
{
|
||
struct type *type = ada_check_typedef (ada_get_base_type (type0));
|
||
char *type_name = decoded_type_name (type0);
|
||
int is_var_decl = (varstring != NULL && varstring[0] != '\0');
|
||
|
||
if (type == NULL)
|
||
{
|
||
if (is_var_decl)
|
||
fprintf_filtered (stream, "%.*s: ",
|
||
ada_name_prefix_len (varstring), varstring);
|
||
fprintf_filtered (stream, "<null type?>");
|
||
return;
|
||
}
|
||
|
||
if (show > 0)
|
||
type = ada_check_typedef (type);
|
||
|
||
if (is_var_decl && TYPE_CODE (type) != TYPE_CODE_FUNC)
|
||
fprintf_filtered (stream, "%.*s: ",
|
||
ada_name_prefix_len (varstring), varstring);
|
||
|
||
if (type_name != NULL && show <= 0 && !ada_is_aligner_type (type))
|
||
{
|
||
fprintf_filtered (stream, "%.*s",
|
||
ada_name_prefix_len (type_name), type_name);
|
||
return;
|
||
}
|
||
|
||
if (ada_is_aligner_type (type))
|
||
ada_print_type (ada_aligned_type (type), "", stream, show, level, flags);
|
||
else if (ada_is_constrained_packed_array_type (type)
|
||
&& TYPE_CODE (type) != TYPE_CODE_PTR)
|
||
print_array_type (type, stream, show, level, flags);
|
||
else
|
||
switch (TYPE_CODE (type))
|
||
{
|
||
default:
|
||
fprintf_filtered (stream, "<");
|
||
c_print_type (type, "", stream, show, level, flags);
|
||
fprintf_filtered (stream, ">");
|
||
break;
|
||
case TYPE_CODE_PTR:
|
||
case TYPE_CODE_TYPEDEF:
|
||
fprintf_filtered (stream, "access ");
|
||
ada_print_type (TYPE_TARGET_TYPE (type), "", stream, show, level,
|
||
flags);
|
||
break;
|
||
case TYPE_CODE_REF:
|
||
fprintf_filtered (stream, "<ref> ");
|
||
ada_print_type (TYPE_TARGET_TYPE (type), "", stream, show, level,
|
||
flags);
|
||
break;
|
||
case TYPE_CODE_ARRAY:
|
||
print_array_type (type, stream, show, level, flags);
|
||
break;
|
||
case TYPE_CODE_BOOL:
|
||
fprintf_filtered (stream, "(false, true)");
|
||
break;
|
||
case TYPE_CODE_INT:
|
||
if (ada_is_fixed_point_type (type))
|
||
print_fixed_point_type (type, stream);
|
||
else
|
||
{
|
||
const char *name = ada_type_name (type);
|
||
|
||
if (!ada_is_range_type_name (name))
|
||
fprintf_filtered (stream, _("<%d-byte integer>"),
|
||
TYPE_LENGTH (type));
|
||
else
|
||
{
|
||
fprintf_filtered (stream, "range ");
|
||
print_range_type (type, stream, 1 /* bounds_prefered_p */);
|
||
}
|
||
}
|
||
break;
|
||
case TYPE_CODE_RANGE:
|
||
if (ada_is_fixed_point_type (type))
|
||
print_fixed_point_type (type, stream);
|
||
else if (ada_is_modular_type (type))
|
||
fprintf_filtered (stream, "mod %s",
|
||
int_string (ada_modulus (type), 10, 0, 0, 1));
|
||
else
|
||
{
|
||
fprintf_filtered (stream, "range ");
|
||
print_range (type, stream, 1 /* bounds_prefered_p */);
|
||
}
|
||
break;
|
||
case TYPE_CODE_FLT:
|
||
fprintf_filtered (stream, _("<%d-byte float>"), TYPE_LENGTH (type));
|
||
break;
|
||
case TYPE_CODE_ENUM:
|
||
if (show < 0)
|
||
fprintf_filtered (stream, "(...)");
|
||
else
|
||
print_enum_type (type, stream);
|
||
break;
|
||
case TYPE_CODE_STRUCT:
|
||
if (ada_is_array_descriptor_type (type))
|
||
print_array_type (type, stream, show, level, flags);
|
||
else if (ada_is_bogus_array_descriptor (type))
|
||
fprintf_filtered (stream,
|
||
_("array (?) of ? (<mal-formed descriptor>)"));
|
||
else
|
||
print_record_type (type, stream, show, level, flags);
|
||
break;
|
||
case TYPE_CODE_UNION:
|
||
print_unchecked_union_type (type, stream, show, level, flags);
|
||
break;
|
||
case TYPE_CODE_FUNC:
|
||
print_func_type (type, stream, varstring, flags);
|
||
break;
|
||
}
|
||
}
|
||
|
||
/* Implement the la_print_typedef language method for Ada. */
|
||
|
||
void
|
||
ada_print_typedef (struct type *type, struct symbol *new_symbol,
|
||
struct ui_file *stream)
|
||
{
|
||
type = ada_check_typedef (type);
|
||
ada_print_type (type, "", stream, 0, 0, &type_print_raw_options);
|
||
fprintf_filtered (stream, "\n");
|
||
}
|