Commit Graph

228 Commits

Author SHA1 Message Date
Ed Hartnett
d994a28139 allocating storage for var hdf5 info 2018-11-13 06:26:36 -07:00
Ed Hartnett
825047b8f6 rest of moving HDF5 specific group info to libhdf5 2018-11-12 14:04:17 -07:00
Ed Hartnett
79c2f843d8 rest of hdf5 specific group changes 2018-11-12 13:57:04 -07:00
Ed Hartnett
971b16cdc8 hdf5 specific changes for hdf5open.c 2018-11-12 13:51:08 -07:00
Ed Hartnett
766f1d0195 hdf5-specific changes for hdf5grp.c 2018-11-12 13:43:43 -07:00
Ed Hartnett
902ee64ff0 hdf5-specific group stuff for hdf5attr.c 2018-11-12 13:38:11 -07:00
Ed Hartnett
261c7fb8d6 hdf5-specific group info for nc4info.c 2018-11-12 13:06:15 -07:00
Ed Hartnett
9c3a8cb2c5 hdf5-specific group stuff in hdf5var.c 2018-11-12 13:01:54 -07:00
Ed Hartnett
5c7fc1fb55 HDF5-specific group changes 2018-11-12 12:52:24 -07:00
Ed Hartnett
e74009c708 HDF5-specific group changes 2018-11-12 12:32:21 -07:00
Ed Hartnett
c59d6930f0 HDF5-specific group changes 2018-11-12 12:12:56 -07:00
Ed Hartnett
65bde4878b HDF5-specific group changes 2018-11-12 11:03:00 -07:00
Ed Hartnett
d6c5e1f127 HDF5-specific group changes 2018-11-12 10:48:24 -07:00
Ed Hartnett
17552b7eb3 HDF5-specific group changes 2018-11-12 10:37:06 -07:00
Ed Hartnett
6b75bb483b allocating HDF5-specific group info struct for root group 2018-11-12 09:02:46 -07:00
Ed Hartnett
9364157a9b allocating and freeing memory for HDF5-specific group info 2018-11-12 08:11:06 -07:00
Ed Hartnett
7534cb24fe cleanup 2018-11-08 11:12:38 -07:00
Ed Hartnett
e88b98daaa moving hdf5_objid field to hdf5-specific dim info 2018-11-08 11:05:20 -07:00
Ed Hartnett
856e4ead03 moved hdf_dimscaleid to hdf5-specific dim info 2018-11-08 10:55:21 -07:00
Ed Hartnett
cc02ec3aa3 more use of HDF5-specific dim info 2018-11-08 10:48:02 -07:00
Ed Hartnett
1dd9004a11 more use of HDF5-specific dim info 2018-11-08 10:43:36 -07:00
Ed Hartnett
7ab71a69bb more use of HDF5-specific dim info 2018-11-08 10:40:39 -07:00
Ed Hartnett
30da28edff using hdf5-specific dim info 2018-11-08 10:37:32 -07:00
Ed Hartnett
0ef97a84f1 cleanup 2018-11-08 10:27:34 -07:00
Ed Hartnett
bff06f2c4d starting to use HDF5-specific dim info 2018-11-08 10:22:37 -07:00
Ed Hartnett
dab52a97ee more looking up of HDF5 dim info 2018-11-08 10:16:55 -07:00
Ed Hartnett
92dc85b802 continuing to set hdf5-specific dim info struct 2018-11-08 10:11:46 -07:00
Ed Hartnett
b85eec1bb8 starting to set hdf5-specific dim info struct 2018-11-08 10:07:59 -07:00
Ed Hartnett
e89ad03b24 more looking up HDF5-specific dim info 2018-11-08 10:02:17 -07:00
Ed Hartnett
3e80723525 continuing to find HDF5 specific dim info 2018-11-08 09:20:01 -07:00
Ed Hartnett
008d4ee796 starting to find HDF5 specific dim info 2018-11-08 08:57:44 -07:00
Ed Hartnett
9373989931 allocating and freeing space for HDF5-specific dim info 2018-11-08 08:51:46 -07:00
Ed Hartnett
8d31f5b806 clean up 2018-11-07 14:23:55 -07:00
Ed Hartnett
6f4b4ac80d moving attribute HDF5 stuff to libhdf5 2018-11-07 14:21:57 -07:00
Ed Hartnett
11d725facc allocating and freeing memory for hdf5-specific attribute info 2018-11-07 13:45:51 -07:00
Ed Hartnett
5f36a3b425 merged master 2018-11-02 10:00:53 -06:00
Dennis Heimbigner
245961de00 re: github issues
https://github.com/Unidata/netcdf-c/issues/1168
    https://github.com/Unidata/netcdf-c/issues/1163
    https://github.com/Unidata/netcdf-c/issues/1162

This PR partially fixes memory leaks in the netcdf-c library,
in the ncdump utility, and in some test cases.

The netcdf-c library now runs memory clean with the assumption
that the --disable-utilities option is used. The primary remaining
problem is ncgen. Once that is fixed, I believe the netcdf-c library
will run memory clean with no limitations.

Notes
-----------
1. Memory checking was performed using gcc -fsanitize=address.
   Valgrind-based testing has yet to be performed.
2. The pnetcdf, hdf4, and examples code has not been tested.

Misc. Non-leak changes
1. Make tst_diskless2 only run when netcdf4 is enabled (issue 1162)
2. Fix CmakeLists.txt to turn off logging if ENABLE_NETCDF_4 is OFF
3. Isolated all my debug scripts into a single top-level directory
   called debug
4. Fix some USE_NETCDF4 dependencies in nc_test and nc_test4 Makefile.am
2018-10-30 20:48:12 -06:00
Ed Hartnett
c2cb3379f4 doing HDF5 closing in libhdf5 2018-10-23 06:59:22 -06:00
Ed Hartnett
35cfaefc0c closing HDF5 objects separately 2018-10-23 05:39:00 -06:00
Ed Hartnett
452f75fadd commented out some tests 2018-10-22 15:04:44 -06:00
Ed Hartnett
958826d0af merged ejh_mem_check 2018-10-22 13:29:46 -06:00
Dennis Heimbigner
e40eb2e950 switch 2018-10-19 11:11:36 -06:00
Ed Hartnett
5cbde66781 now closing att HDF5 typeid in HDF5 close code 2018-10-18 03:43:53 -06:00
Ed Hartnett
695e295734 now calling recursive function to close HDF5 objects in file 2018-10-18 03:29:21 -06:00
Ed Hartnett
fa86d3c488 continuing to separate hdf5/libsrc4 file close code 2018-10-18 03:20:39 -06:00
Ed Hartnett
c90ab24b48 moving towards separating HDF5 file close from netcdf4 file close 2018-10-18 03:17:38 -06:00
Dennis Heimbigner
979873f81d Fix provenance memory leak 2018-10-17 11:43:14 -06:00
Dennis Heimbigner
4636584d5b Revert/Improve nc_create + NC_DISKLESS behavior
re: https://github.com/Unidata/netcdf-c/issues/1154

Inadvertently, the behavior of NC_DISKLESS with nc_create() was
changed in release 4.6.1. Previously, the NC_WRITE flag needed
to be explicitly used with NC_DISKLESS in order to cause the
created file to be persisted to disk.

Additional analyis indicated that the current NC_DISKLESS
implementation was seriously flawed.

This PR attempts to clean up and regularize the situation with
respect to NC_DISKLESS control. One important aspect of diskless
operation is that there are two different notions of write.

1. The file is read-write vs read-only when using the netcdf API.
2. The file is persisted or not to disk at nc_close().

Previously, these two were conflated. The rules now are
as follows.

1. NC_DISKLESS + NC_WRITE means that the file is read/write using the netcdf API
2. NC_DISKLESS + NC_PERSIST means that the file is persisted to a disk file at nc_close.
3. NC_DISKLESS + NC_PERSIST + NC_WRITE means both 1 and 2.

The NC_PERSIST flag is new and takes over the obsolete NC_MPIPOSIX flag.
NC_MPIPOSIX is still defined, but is now an alias for the NC_MPIIO flag.

It is also now the case that for netcdf-4, NC_DISKLESS is independent
of NC_INMEMORY and in fact it is an error to specify both flags
simultaneously.

Finally, the MMAP code was fixed to use NC_PERSIST as well.
Also marked MMAP as deprecated.

Also added a test case to test various combinations of NC_DISKLESS,
NC_PERSIST, and NC_WRITE.

This PR affects a number of files and especially test cases
that used NC_DISKLESS.

Misc. Unrelated fixes
1. fixed some warnings in ncdump/dumplib.c
2018-10-10 13:32:17 -06:00
Dennis Heimbigner
a5a34f6aba
Merge branch 'master' into nc_mpiio_nc_mpiposix 2018-10-06 13:33:55 -06:00
Wei-keng Liao
0ed70756cc Ignore flags NC_MPIIO and NC_MPIPOSIX. 2018-09-22 20:22:34 -05:00