Commit Graph

947 Commits

Author SHA1 Message Date
Edward Hartnett
a6fbc3eea2 fix problem with pre-enddef call to nc_inq_var_szip() 2020-02-04 07:11:44 -07:00
Edward Hartnett
bf3407ce67 more tests 2020-02-04 05:38:15 -07:00
Edward Hartnett
e363bd59c7 testing nc_def_var_szip when szip has not been built into HDF5 2020-02-04 05:33:52 -07:00
Edward Hartnett
21a204c4e8 start using par filters in HDF5-1.10.3 2020-01-17 14:01:07 -07:00
Edward Hartnett
b9823d2658 fixed shell test for par filters 2020-01-17 12:13:45 -07:00
Edward Hartnett
3da6304f6a fixing running of parallel filter tests only when parallel filters are present 2020-01-17 11:33:58 -07:00
Edward Hartnett
86a9835f12 checking sziped data in parallel I/O test 2020-01-06 10:23:59 -07:00
Edward Hartnett
58abb5adee parallel szip test 2020-01-06 10:23:07 -07:00
Edward Hartnett
c8135722ac parallel szip test 2020-01-06 10:22:45 -07:00
Edward Hartnett
a8a523f1aa parallel szip test 2020-01-06 10:21:00 -07:00
Edward Hartnett
a0396a862b parallel szip test 2020-01-06 10:17:28 -07:00
Edward Hartnett
bb7ffe69d9 starting to add parallel tests 2020-01-06 09:08:27 -07:00
Edward Hartnett
7e332690d2 more szip testing 2020-01-06 08:04:27 -07:00
Edward Hartnett
9c8719bd00 more szip tests 2020-01-06 08:01:28 -07:00
Edward Hartnett
65e5897533 whitespace clean up 2020-01-06 07:58:11 -07:00
Edward Hartnett
8a341372ea clean up 2020-01-06 07:57:50 -07:00
Edward Hartnett
1c3edc503d now checking for proper error from nc_def_var_szip() when szip is not present 2020-01-06 07:53:31 -07:00
Edward Hartnett
c47b404531 added test for when szip is missing from HDF5 2020-01-06 07:50:02 -07:00
Edward Hartnett
808a0e2be9 merged ejh_parallel_zlib 2020-01-02 14:25:31 -07:00
Ward Fisher
8f2be58d95
Merge pull request #1566 from NetCDF-World-Domination-Council/ejh_unlim_dims
Fix problems with read past end of dataset but within dimension length for vars with multiple unlimited dimensions
2019-12-23 15:08:56 -07:00
Edward Hartnett
af0b28ffd5 removed unneeded delete 2019-12-20 14:05:57 -07:00
Edward Hartnett
c4a8671a0d fixed comments 2019-12-20 14:04:57 -07:00
Edward Hartnett
61477db387 fixed comments 2019-12-20 14:03:48 -07:00
Edward Hartnett
9ded9b1dfd restored tst_mode tests 2019-12-20 14:00:59 -07:00
Edward Hartnett
680e44f628 changed name of macro 2019-12-20 13:58:01 -07:00
Edward Hartnett
c16e7ef3ce more test development 2019-12-20 12:52:38 -07:00
Edward Hartnett
522705423d more test development 2019-12-20 12:51:36 -07:00
Edward Hartnett
909212454a more testing 2019-12-20 12:43:52 -07:00
Edward Hartnett
8e660e14fa more tests 2019-12-20 11:48:40 -07:00
Edward Hartnett
e6ded1a8ec more testing 2019-12-20 11:42:54 -07:00
Edward Hartnett
995cfdad96 merged master 2019-12-20 11:16:11 -07:00
Edward Hartnett
8681b0d241 more documentaiton 2019-12-20 07:10:13 -07:00
Edward Hartnett
4b7f839666 swtich to collective access when filters are applied 2019-12-20 07:00:12 -07:00
Edward Hartnett
f980743be1 now also can set fletcher32 in parallel I/O writes 2019-12-20 06:48:31 -07:00
Edward Hartnett
0c6641eb98 adjusted tst_parallel_zlib2 to HDF5 version flag 2019-12-20 06:46:13 -07:00
Edward Hartnett
663540b3a7 adjusted tst_parallel_zlib to HDF5 version flag 2019-12-20 06:44:14 -07:00
Edward Hartnett
6d2d92ddec added new tests to cmake, also additional test development 2019-12-20 06:30:23 -07:00
Edward Hartnett
d534b1298a adding another zlib parallel I/O test 2019-12-20 05:28:20 -07:00
Edward Hartnett
2a415dba7c stop running tst_mode test 2019-12-20 05:22:53 -07:00
Ward Fisher
6c75e97764
Merge pull request #1570 from NOAA-GSD/ejh_compact
enable compact storage for netcdf-4 vars
2019-12-19 16:47:05 -07:00
Edward Hartnett
cde58d23c1 allowing parallel write of gzipped data, plus added test 2019-12-19 09:20:20 -07:00
Edward Hartnett
3e00967879 allow parallel writes to use zlib 2019-12-19 09:19:23 -07:00
Edward Hartnett
fb4a209f2a more testing for compact vars 2019-12-16 09:44:31 -07:00
Edward Hartnett
e75c248478 more testing for compact vars 2019-12-16 09:42:54 -07:00
Edward Hartnett
66a2b4c05e more testing for compact vars 2019-12-16 09:37:54 -07:00
Edward Hartnett
90324dfd10 further test development for compact storage 2019-12-16 09:31:11 -07:00
Edward Hartnett
89b8981c86 got compact storage test working 2019-12-04 09:12:53 -07:00
Edward Hartnett
06896f432d got compact storage test working 2019-12-04 08:49:37 -07:00
Edward Hartnett
85994847b2 more test development 2019-12-04 08:06:55 -07:00
Edward Hartnett
1a1f537c26 more test development 2019-12-04 08:05:30 -07:00
Edward Hartnett
1a665b6d80 adding test for compact storage 2019-12-04 08:01:56 -07:00
Edward Hartnett
bb1f5e1637 whitespace cleanup of test 2019-12-04 07:55:54 -07:00
Edward Hartnett
645b8566cd more tests 2019-12-01 15:05:43 -07:00
Edward Hartnett
e52a74520e tests and fix for multiple unlimited dim bug 2019-12-01 15:05:09 -07:00
Edward Hartnett
c1756e949e merged ejh_remove_unrelax 2019-12-01 14:30:37 -07:00
Edward Hartnett
c0a6da7757 changed test filename to one that will be cleaned 2019-12-01 08:47:46 -07:00
Edward Hartnett
ddf986665b comment to explain 2019-12-01 07:35:41 -07:00
Edward Hartnett
f86ca19ee3 added test for opening HDF5 file without creation ordering 2019-12-01 07:30:32 -07:00
edwardhartnett
8f4d5e6a5b turned off cmake build in travis 2019-11-17 15:41:24 -07:00
Ward Fisher
af8f9ad2cf
Merge pull request #1523 from NetCDF-World-Domination-Council/ejh_udf
User-defined formats must come first in NC_infermodel, plus test
2019-11-15 16:40:07 -07:00
Ward Fisher
923d4ccbff
Merge pull request #1530 from NetCDF-World-Domination-Council/ejh_endianness
now testing that endianness can only be set on atomic ints and floats
2019-11-15 15:27:35 -07:00
edwardhartnett
965da1de01 now testing that endianness can only be set on atomic ints and floats 2019-11-15 11:10:10 -07:00
edwardhartnett
b844f8f8f6 more testing 2019-11-15 09:26:37 -07:00
edwardhartnett
35e210ac29 test cleanup 2019-11-15 09:21:54 -07:00
edwardhartnett
0017f3e069 test cleanup 2019-11-15 09:21:11 -07:00
edwardhartnett
7f41db9cb3 test cleanup 2019-11-15 09:20:26 -07:00
edwardhartnett
8083b3596e fixed problem of unlim dim and var sharing the same name but not being related 2019-11-15 09:18:42 -07:00
Ward Fisher
1a6351dab2
Merge pull request #1521 from ckhroulev/netcdf4-repeated-attribute-modification
Improve the fix for #350 included in #1119
2019-11-14 17:01:09 -07:00
Ward Fisher
2462cda15e
Merge pull request #1525 from NetCDF-World-Domination-Council/ejh_anon_dims
If HDF5 dataset has multiple anonymous dimensions of the same size, assume they are different dimensions
2019-11-14 16:58:24 -07:00
edwardhartnett
d0b069cf70 cleaned up test handling of strings 2019-11-14 09:39:10 -07:00
edwardhartnett
3e45fa1497 another test for two anon dimensions of same size used in same HDF5 var 2019-11-14 07:23:09 -07:00
edwardhartnett
d73611de73 now handle two anon dimensions of same size used in same HDF5 var 2019-11-14 06:54:22 -07:00
edwardhartnett
cebe84157b adding test for anonymous dims in HDF5 file 2019-11-13 12:51:34 -07:00
edwardhartnett
b9f57b2b5f now UDF mode flags take priority over NC_NETCDF4 2019-11-13 12:13:33 -07:00
Constantine Khrulev
bb45db457e Add regression tests modifying attributes 2^16 times
This commit adds three new tests:

- a test documenting the limitation originally reported in #350 (in
  general modifying an attribute about 2^16 times makes it impossible
  to modify this file in ways requiring nc_redef() and nc_enddef() calls).
- a test ensuring that a scalar attribute can be modified 2^16 times
  as long as its type and size remain the same
- a test ensuring that a text attribute can be modified 2^16 times as
  long as its size remains the same

This version uses the nc_redef(), nc_put_att_...(), nc_enddef()
sequence. One could also use nc_open(), nc_put_att_...(), nc_close()
but that would make these tests significantly slower.
2019-11-13 10:08:31 -09:00
edwardhartnett
0bbe91e438 udf must take priority in NC_infermodel 2019-11-13 12:07:33 -07:00
Dennis Heimbigner
30cfd348e6 Remove debug output 2019-11-06 09:33:14 -07:00
Dennis Heimbigner
9e016b85aa Add test cases 2019-11-03 12:03:13 -07:00
Dennis Heimbigner
f1506d552e Change (again), and hopefully simplify, the file model inference algorithm.
* For URL paths, the new approach essentially centralizes all information
  in the URL into the "#mode=" fragment key and uses that value
  to determine the dispatcher for (most) URLs.

* The new approach has the following steps:

  1. canonicalize the path if it is a URL.
  2. use the mode= fragment key to determine the dispatcher
  3. if dispatcher still not determined, then use the mode flags
     argument to nc_open/nc_create to determine the dispatcher.
  4. if the path points to something readable, attempt to read the
     magic number at the front, and use that to determine the dispatcher.
     this case may override all previous cases.

* Misc changes.

  1. Update documentation
  2. Moved some unit tests from libdispatch to unit_test directory.
  3. Fixed use of wrong #ifdef macro in test_filter_reg.c
     [I think this may fix an previously reported esupport query].
2019-09-29 12:59:28 -06:00
Greg Sjaardema
56c0d5cf8a Spelling fixes 2019-09-18 08:03:01 -06:00
Ward Fisher
9a92201c94 Wiring unit test directory into cmake-based builds. 2019-08-21 14:50:09 -06:00
edwardhartnett
2077729abc removed base_pe functions from dispatch table 2019-08-15 06:51:06 -06:00
edwardhartnett
0b00f4b5eb more work on tst_parallel5.c 2019-08-14 08:50:09 -06:00
edwardhartnett
c042ea5d23 added parallel I/O test for NC_BYTE scalar 2019-08-14 08:22:35 -06:00
edwardhartnett
170c5b0901 removed NC from open in dispatch table 2019-08-01 14:30:20 -06:00
Even Rouault
9d1331ddc9
tst_bug1442: improve test to retrieve a mix of valid values and values at fillvalue 2019-07-18 03:05:20 +02:00
Even Rouault
3f268203db
tst_bug1442.c: address review comments 2019-07-18 02:41:06 +02:00
Even Rouault
16babd3e89
nc_test4/Makefile.am: add tst_bug1442 2019-07-18 02:38:28 +02:00
Even Rouault
0c7be1d278
Add test case for bugfix of #1442 2019-07-18 02:23:43 +02:00
Ed Hartnett
d5816bb13e using netcdf_dispatch.h in tst_udf.c 2019-07-05 05:24:29 -06:00
Ward Fisher
d6a3944199
Merge pull request #1409 from Unidata/nccopydefault.dmh
Nccopy was overriding default chunking when it should not.
2019-05-29 15:26:09 -06:00
Dennis Heimbigner
06498ff16a various fixes 2019-05-23 16:35:03 -06:00
Ed Hartnett
73a433240c fixed warnings in tst_large2.c 2019-05-21 15:24:56 -06:00
Ed Hartnett
d858313d2e fixed warnings in tst_udf.c 2019-05-21 15:23:23 -06:00
Ed Hartnett
9fd7c06f16 fixed warnings in tst_rename2 2019-05-21 15:16:35 -06:00
Ed Hartnett
f35eea13d7 fixed warnings in tst_rename3 2019-05-21 15:15:25 -06:00
Ed Hartnett
fd29e2e608 uncommented test 2019-05-12 08:51:02 -06:00
Ward Fisher
103e5a3ff5
Merge pull request #1385 from Unidata/mmapfix.dmh
Fix cmake wrt mmap
2019-05-02 11:09:12 -06:00
Ward Fisher
44fce0904e
Merge pull request #1387 from Unidata/fixffilter.dmh
Minor config.h changes to support filters in Fortran
2019-05-01 14:44:53 -06:00
Ward Fisher
5410967b00 Merge branch 'master' into addfilter.dmh 2019-04-30 14:51:25 -06:00
Dennis Heimbigner
62feacee00 missing cmake file 2019-04-29 20:55:28 -06:00
Dennis Heimbigner
c9d16d82d6 Fix cmake X mmap
supercede PR: https://github.com/Unidata/netcdf-c/pull/1384

Since we have an mmap user, undeprecate it and make sure
it works. Other changes:

* fix test cases to work with make -j
* fix exposed ncgen error.
2019-04-19 20:32:26 -06:00
Ed Hartnett
d4ab03ea23 merged upstream/master 2019-03-26 11:29:01 -06:00
Ward Fisher
85b94a4244
Merge pull request #1364 from NetCDF-World-Domination-Council/ejh_non_contoversial
non controversial parts of last PR
2019-03-26 11:14:33 -06:00
Dennis Heimbigner
8d0bced60d Allow in-line definition of filters
Priority: Low

re: issue https://github.com/Unidata/netcdf-c/issues/1329

HDF5 has the ability to programmatically define new filters,
as opposed to using HDF5_PLUGIN_PATH env variable.
This PR adds support for that feature.
Not clear how useful this is, though.
See docs/filters.md for details.
2019-03-21 11:33:27 -06:00
Ward Fisher
a25236eac4
Merge pull request #1348 from Unidata/pluginst.dmh
Support installation of the bzip2 plugin under name libh5bzip2.so
2019-03-20 15:42:06 -06:00
Ed Hartnett
54eb9dad77
Merge branch 'master' into ejh_non_contoversial 2019-03-19 14:16:52 -06:00
Ward Fisher
e2b31ffae4
Merge branch 'master' into byterange.dmh 2019-03-19 12:05:44 -06:00
Ed Hartnett
d709624174 now run tst_put_vars_two_unlim_dim 2019-03-18 11:16:44 -06:00
Ed Hartnett
8799fff54b cleanup Makefile.am 2019-03-18 11:09:44 -06:00
Ed Hartnett
06c9f681f4 got distcheck target working 2019-03-18 10:15:18 -06:00
Ed Hartnett
94d9cd7c8f don't allow benchmarks for classic only builds 2019-03-18 08:40:18 -06:00
Ed Hartnett
e3c628482e moving metadata performance tests to nc_perf 2019-03-18 08:30:14 -06:00
Ed Hartnett
51b8ba10b4 attempting to get cmake build working with new nc_perf directory 2019-03-18 08:09:46 -06:00
Ed Hartnett
cf7126b294 fixing distcheck issues 2019-03-17 11:17:29 -06:00
Ed Hartnett
e48bd5e4c2 removed benchmarking stuff from nc_test4/Makefile.am 2019-03-17 10:57:45 -06:00
Ed Hartnett
286dddac4b moved many other benchmark tests to nc_perf 2019-03-17 10:57:00 -06:00
Ed Hartnett
95a0b8ddfb moved other benchmark programs to nc_perf 2019-03-17 08:30:06 -06:00
Ed Hartnett
e391a9b41f moved bm_file.c 2019-03-17 08:19:17 -06:00
Ed Hartnett
a3ca0e4333 move benchmark test script to nc_perf 2019-03-17 08:15:29 -06:00
Ed Hartnett
16c68218de move parallel benchmark test script to nc_perf 2019-03-17 08:14:27 -06:00
Ed Hartnett
73354a9862 moved tst_create_files from nc_test4 to nc_perf 2019-03-17 08:11:08 -06:00
Ed Hartnett
7902615b1c adding nc_perf directory 2019-03-17 08:03:27 -06:00
Ed Hartnett
f91e246699 turned off some ncdump test output 2019-03-12 09:26:15 -06:00
Ed Hartnett
c89fad34e0 only run some benchmark tests if utilities have been built 2019-03-12 09:23:49 -06:00
Dennis Heimbigner
15e4fdcbb4 Support installation of the bzip2 plugin under name libh5bzip2.so
re: https://github.com/Unidata/netcdf-c/issues/1347

It turns out that the plugin libraries (bzip2 and misc) were
being installed as part of 'make installed'. This was not intended
behavior. But after some discussion in the above issue, it was decided
to install the bzip2 plugin. However, in order to avoid naming conflicts,
the plugin is installed under the name 'libh5bzip2.so'.
Note that this is automake behavior only; the install does not
(yet) occur using cmake.

Misc. unrelated changes
-----------------------
1. turn off some debug output in ncdump/Makefile.am
2019-02-27 15:22:46 -07:00
Dennis Heimbigner
6d86aab022 Modify tst_filterparser to use baseline array
re: https://github.com/Unidata/netcdf-c/pull/1337

tst_filterparser was creating but not using the baseline
array. Instead, it did the same thing inline. Fixed code
to use baseline array.
2019-02-25 14:31:54 -07:00
Dennis Heimbigner
4c1ed0144b Fix nc_test4/tst_filter.sh for big endian
re: issue https://github.com/Unidata/netcdf-c/issues/1338

Changes:

1. nc_test4/tst_filter.sh + nc_test4/ref_filteredvv.cdl --
   properly suppress _Endianness attribute
2. fix some warnings
2019-02-24 22:20:01 -07:00
Dennis Heimbigner
0c59e13bf7 Master merge, conflict resolution, cleanup 2019-02-24 16:54:13 -07:00
Dennis Heimbigner
45a8a265b8 master merge 2019-02-23 17:14:12 -07:00
Ward Fisher
404f87b8c2 Turned of filterparser test when building static library. 2019-02-20 15:11:06 -07:00
Ward Fisher
53ba2ff316
Merge pull request #1311 from Unidata/filterexpr.dmh
Extend nccopy -F option syntax.
2019-02-15 15:16:25 -07:00
Dennis Heimbigner
a1f080db19
Merge branch 'master' into byterange.dmh 2019-02-13 14:45:53 -07:00
Ward Fisher
e32e0b1eb2
Merge branch 'master' into filterexpr.dmh 2019-02-12 09:48:03 -07:00
Ed Hartnett
909402ccd6
Merge branch 'master' into ejh_more_rename_woes_3 2019-02-12 05:49:04 -07:00
Dennis Heimbigner
5f0fdd7e5d Update Makefile.am to use ref_ filter files 2019-02-11 16:07:03 -07:00
Dennis Heimbigner
e932655888 Renamed the baseline files for tst_filter.sh 2019-02-09 15:14:13 -07:00
Dennis Heimbigner
a6b04c0c66 Extend nccopy -F option syntax.
A user suggested that the nccopy -F option
syntax should be extended to support specification
of multiple (or all) variables in a single -F option.

The new syntax allows:

1. '*' as the name of the variable; this means apply the
   filter to all variables in the data set.
2. *var1|var2|...* as the variable name to indicate that the filter
   should be applied to the multiple specified variables.
2019-02-08 18:48:17 -07:00
Ward Fisher
4c6d56abc0
Merge branch 'master' into ejh_more_rename_woes_3 2019-02-07 14:28:46 -07:00
Ward Fisher
3a14251dd9
Merge branch 'master' into bigend.dmh 2019-02-07 14:28:39 -07:00
Ward Fisher
1fde39c8d7
Merge branch 'master' into byterange.dmh 2019-02-07 14:28:23 -07:00
Ward Fisher
6511599a6d Corrected issue in support of https://github.com/Unidata/netcdf-c/issues/1310 2019-02-07 12:43:07 -07:00
Ed Hartnett
5d908a0bbb now preserve order of varids after a var rename 2019-02-03 06:56:03 -07:00
Ed Hartnett
275e116f49 added test to demonstrate out of order varids after rename 2019-02-03 06:28:04 -07:00
Ed Hartnett
6728fceab6 uncommented a bunch of rename tests 2019-02-02 05:55:18 -07:00
Ed Hartnett
1dd76c996e added tst_rename3.c for more rename testing 2019-02-02 05:53:45 -07:00
Ed Hartnett
c8cbff4728 added test (commented out) in tst_rename2.c which fails 2019-02-02 04:00:34 -07:00
Dennis Heimbigner
6723a38d90 remove some debug statements 2019-02-01 14:39:30 -07:00
Dennis Heimbigner
65e9a46d84 ckp 2019-02-01 12:00:12 -07:00
Dennis Heimbigner
8714066b18 Fix errors when building on big-endian machine
re: issue https://github.com/Unidata/netcdf-c/issues/1278
re: issue https://github.com/Unidata/netcdf-c/issues/876
re: issue https://github.com/Unidata/netcdf-c/issues/806

* Major change to the handling of 8-byte parameters for nc_def_var_filter.
  The old code was not well thought out.
  * The new algorithm is documented in docs/filters.md.
  * Added new utility file plugins/H5Zutil.c to support
  * Modified plugins/H5Zmisc.c to use new algorithm
  the new algorithm.
  * Renamed include/ncfilter.h to include/netcdf_filter.h
    and made it an installed header so clients can access the
    new algorithm utility.
  * Fixed nc_test4/tst_filterparser.c and nc_test4/test_filter_misc.c
    to use the new algorithm
* libdap4/ fixes:
  * d4swap.c has an error in the endian pre-processing such
    that record counts were not being swapped correctly.
  * d4data.c had an error in that checksums were being computed
    after endian swapping rather than before.
* ocinitialize() was never being called, so xxdr bigendian handling
  was never set correctly.
  * Required adding debug statements to occompile
* Found and fixed memory leak in ncdump.c

Not tested:
* HDF4
* Pnetcdf
* parallel HDF5
2019-01-31 21:13:06 -07:00