summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* [svn-r15820] Added a design document describing the design details for the ↵inactive/modify_dtype_filterRaymond Lu2008-10-081-0/+0
| | | | | | filter of modifying the dataset's datatype. For future reference.
* [svn-r15765] This is the second checkin for the filter of modifying ↵Raymond Lu2008-10-0333-346/+3064
| | | | | | | | | | | | | | | | | dataset's datatype. There're a lot of changes since the last checkin. Basically, I put a skipped list in the filter to keep track of the old (current) datatype for the chunk when the data chunk is read into the memory. When the datatype is changed, the filter will decrement the reference count for the old datatype. Another significant change is putting a B-tree in the filter to record the reference count of some datatype from the data chunks. When the dataset is deleted, the filter will decrement this many reference count for the datatype all at once to save time. This project is wrapped up at this point temporarily because of the flaws in the design. Quincey will talk to the people in Chicago about redesigning it. I've tested it on smirom, kagiso, and linew.
* [svn-r14400] Added some checking for the validity of reference datatype and ↵Raymond Lu2008-01-114-2/+436
| | | | | | | | | the test for it. Mistakenly, H5Zdtype_modify.c wasn't committed in last round. Tested on smirom, kagiso, and linew.
* [svn-r14351] Some modification of the code for the filter of modifying ↵Raymond Lu2007-12-187-211/+96
| | | | | | dataset's datatype. Tested on smirom, kagiso, and linew.
* [svn-r14349] This is the first commit after branching out the feature of the ↵Raymond Lu2007-12-1797-315/+4437
| | | | | | | | | | | | filter for modifying dataset's datatype. There seems a lot of changes. Basically, this feature allows an application to change the datatypes of datasets. If the dataset is chunked and this filter is enabled, the modification of the data is until new data is written to the file. If the dataset is contiguous or compact, the data is converted immediately when H5Dmodify_dtype is called. I've tested it on 3 platforms. But there'll be some more changes coming. There'll be more thorough test later.
* [svn-r14346] New feature: has_filters_obj function that checks if filters ↵Pedro Vicente Nunes2007-12-145-18/+134
| | | | | | | | | | | | | | | | | | | | match for each individual dataset Following the new feature of h5repack to allow multiple filters for all datasets and the new function has_filters that checks if the repacked file has all the filters requested, I added a new function has_filters_obj that does the same for each dataset. The previous function that checked this only ckecked if the user input filters were in the output dataset. This new function does this but checks if the filters are exactly the same. Currently the behavior of h5repack is to delete all filters that are present in the input file (dataset) and replace them with the requested ones, so they must match exactly. We might consider adding other logical operations, like keep the existing ones. Additionally , the function also checks if the filter parameters match. While doing this I noticed that for the shuffle filter , the values returned do not match and also the same for the N-bit and scale-offset The new function that checks for the filter values fails then, and so I commented the h5repack tests that do this for the N-bit and scale-offset filter (previously for the same bug on the shuffle filter I added special code on the compare filter function but this is temporary until I find the issue) tested: windows, linux, solaris
* [svn-r14345] substituted the symbol CD_VALUES , number of filter client data ↵Pedro Vicente Nunes2007-12-143-6/+8
| | | | | | | | values, for H5Z_COMMON_CD_VALUES that is defined in the library tested: windows, linux, solaris
* [svn-r14344] Maintenance on VMS system: updated testscript for h5diff and ↵Elena Pourmal2007-12-133-299/+73
| | | | | | removed obsolete command file Platforms tested: VMS server
* [svn-r14343] Maintenance for VMS: added more tests to the command file for ↵Elena Pourmal2007-12-133-206/+86
| | | | | | | | h5dump testing modified the file to save output into h5dump_output.txt file removed obsolete test script file check.com update MANIFEST to refelect the change
* [svn-r14342] change the implementation of H5IMmake_image_8bit to use a 2D ↵Pedro Vicente Nunes2007-12-134-10/+1
| | | | | | dataset tested: windows, linux solaris w/fortran
* [svn-r14341] Maintenance for VMS system: VMS C++ compiler expects 'cxx' ↵Elena Pourmal2007-12-131-2/+2
| | | | | | | | | | extensions. Used copy command instead of rename to get the correct names of the C++ source files to avoid errors when build runs second time in the same directory. Platforms tested: VMS server
* [svn-r14340] Maintenance for VMS system:Elena Pourmal2007-12-139-8/+51
| | | | | | | | | | | | | | | | | Skip one of the tests in flush2.c since it looks like VMS system flsuhes the file anyway (like on Winodws with DLLs) Renamed files family*.h5 that are used in backward compatibility testing to have only one '.' in a file name, otherwise VMS system becomes confused and test fails. Cleaned up the only test (Fortran multi-file test) that is failing on VMS (C counterpart passesa). Added new script to facilitate the build process on VMS Updated MANIFEST PLatforms tested: VMS server, kagiso and linew (minor changes)
* [svn-r14339] bug fix: aligned a line in usage that was not alignedPedro Vicente Nunes2007-12-122-2/+2
| | | | tested: linux
* [svn-r14338] Maintenance: Cleaned up system warnings for the install.com scriptElena Pourmal2007-12-111-9/+6
| | | | Platforms tested: VMS server
* [svn-r14337] Maintenance: Added new install.com file to the list of files.Elena Pourmal2007-12-111-0/+1
| | | | Platforms tested: ran bin/chkmanifest script on kagiso
* [svn-r14336] Maintenance/VMS port: added new command file to install C, ↵Elena Pourmal2007-12-116-77/+224
| | | | | | | | Fortran and C++ libraries and corresponding examples; modified some command files to check installation. Platforms tested: VMS server
* [svn-r14335] Corrected the comment on line 219: hard -> softH. Joe Lee2007-12-111-1/+1
|
* [svn-r14334] Description:Quincey Koziol2007-12-105-114/+188
| | | | | | | | | | | | | | | | | | | | | | | | | - Avoid trying to update 1-D dataset's chunk indices (they can't change) - Cache a copy of a dataspace describing a chunk, when doing single element I/O - Keep a 'chunk info' struct around, for single element I/O - Avoid creating a skip list for chunk infos when performing I/O on single element Also, minor formatting cleanups to testing code Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode Linux/64-ia64 2.4 (tg-login3) w/parallel, w/FORTRAN, in production mode
* [svn-r14333] fix typo in commentPedro Vicente Nunes2007-12-101-1/+1
|
* [svn-r14331] Purpose: Update Windows h5ls test scriptScott Wegner2007-12-101-16/+6
| | | | | | | | Details: We no longer need to parse the h5ls output to remove "Modified" tags. This makes it much easier to test on Windows, because we don't need to emulate `sed`. Tested: VS2005 on WinXP
* [svn-r14327] Maintenance for VMS: brought command file up-to-dateElena Pourmal2007-12-061-3/+3
| | | | Platforms tested: VMS server at Boeing.
* [svn-r14326] Description:Quincey Koziol2007-12-0613-208/+299
| | | | | | | | | | | | | | | | | | | | | | | | - Keep skip list for tracking chunks with dataset (instead of creating/ destroying it for each I/O operation) and just delete the skip list nodes. - Avoid computations for normalizing selection offset when offset not set. - Avoid updating object modification time twice during dataset creation. - Avoid updating dataset layout message (and object modification time) until dataset is closed. Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode Linux/64-ia64 2.4 (tg-login3) w/parallel, w/FORTRAN, in production mode
* [svn-r14325] Eliminate unwanted output and remove superfluous sed script. ↵Larry Knox2007-12-061-3/+4
| | | | Tested on kagiso, linew, and smirom.
* [svn-r14324] bug fix: #786 (E3) h5dump refers to the user block as "boot block"Pedro Vicente Nunes2007-12-062-5/+5
| | | | | | replaced usage message with "super block" tested: linux
* [svn-r14320] Added a function that matches the filters on the dataset's DCPL ↵Pedro Vicente Nunes2007-12-045-70/+291
| | | | | | | | and in the requested list Used to verify if the filters requested are present in the output dataset tested: windows, linux, solaris
* [svn-r14319] Purpose: Update error handling in h5ls, uncomment testScott Wegner2007-12-041-30/+34
| | | | | | | | Description: Larry was able to track down the bug in the h5ls test script and change it around to check for an expected error code for each test. With this, we are now able to uncomment the broken test on Windows. Tested: VS2005 on WinXP
* [svn-r14318] Check error code of h5ls call as well as expected output. ↵Larry Knox2007-12-031-25/+30
| | | | Tested kagiso linew smirom.
* [svn-r14317] code maintenance: code cleaning and reorganizing of the ↵Pedro Vicente Nunes2007-12-032-364/+382
| | | | | | h5repack verify routines tested: windows, linux
* [svn-r14316] Purpose: Update test directories in MinGW documentationScott Wegner2007-12-031-1/+3
| | | | | | | | Description: There are many tests that are unsupported on MinGW because of the way it handles certain syntax in shell scripts (particularly '/'). As a result, we list a few tests that users must comment out in various Makefile.in's. This checkin updates the location of h5stat-- it was formerly in ./tools/misc, but has been moved it ./tools/h5stat. Tested: MinGW on WinXP
* [svn-r14315] Purpose: Update Windows h5repack test scriptScott Wegner2007-12-031-1/+22
| | | | | | | | Description: Added new test for long switch names, which was added to the general test script a few days ago. Tested: VS2005 on WinXP
* [svn-r14312] Added support of Albert Cheng2007-11-291-4/+16
| | | | | | | | -fortran: add --enable-fortran -cxx: add --enable-cxx -pp: add --enable-parallel Tested: Kagiso (serial and parallel).
* [svn-r14311] enhancement: change the usage example for multiple filtersPedro Vicente Nunes2007-11-291-2/+2
| | | | | | to note that SHUF should be used before GZIP and that the order specified in command line is the order filters are applied tested: windows, linux
* [svn-r14310] error check in command line: check for a wrong input of the 2 ↵Pedro Vicente Nunes2007-11-291-0/+8
| | | | | | file names being the same tested: windows, linux
* [svn-r14309] bug fix: add a check for the presence of a filter in the call ↵Pedro Vicente Nunes2007-11-291-4/+15
| | | | | | | | to test of long swtich names add a test for multiple global filters tested: linux, solaris
* [svn-r14308] new feature test: add a test for the long switch namesPedro Vicente Nunes2007-11-291-1/+9
| | | | tested: linux, solaris
* [svn-r14307] Description:Quincey Koziol2007-11-299-256/+542
| | | | | | | | | | | | | | | | | | | | | | | | - Extracted common code from H5D_select_read/H5D_select_write into single routine (H5D_select_io) and made H5D_select_read/H5D_select_write "gateway" routines, passing I/O buffer to operate on in "op" struct. - Create optimized pathway in H5D_select_io for I/O operations on single elements. - Make "get linear offset of first element" callback for each type of selection (used in "optimized pathway", above) Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode Linux/64-ia64 2.4 (tg-login3) w/parallel, w/FORTRAN, in production mode
* [svn-r14306] Description:Quincey Koziol2007-11-295-129/+125
| | | | | | | | Add call to reset cached chunk info after initialization of chunk data storage. Tested on: Linux/64 2.4(smirom) w/parallel
* [svn-r14305] updated commentsPedro Vicente Nunes2007-11-291-0/+2
| | | | tested: windows, linux
* [svn-r14304] change feature: rename compression used as %Savings Pedro Vicente Nunes2007-11-291-5/+3
| | | | | | | | | | | now the minus sign shows there was a DECREASE in compression percentage is calculated from per = (b-a)/a; where a= size of dataset before b = size after tested: windows, linux
* [svn-r14303] Purpose: Update h5ls test script on Windows.Scott Wegner2007-11-291-2/+7
| | | | | | | | Description: The h5ls test script was recently updated on Linux. This makes the same changes on Windows. Also, we had to comment out one of our tests, because h5ls returns an error-code that is giving false-positives. Tested: VS2005 on WinXP
* [svn-r14302] Purpose: Update Windows test scripts for recent changesScott Wegner2007-11-292-5/+5
| | | | | | | | | | Description: Minor changes were recently made in the linux test scripts for h5diff and h5repack. This brings Windows up-to-date. Note: I am still working on updates to testh5ls.bat-- there is one test that is failing, but should be ready soon. Tested: VS2005 on WinXP
* [svn-r14301] Purpose: Remove references to deleted source file from Windows ↵Scott Wegner2007-11-296-154/+0
| | | | | | | | | | project files. Description: h5repack_list.c was deleted from the repository recently. This check-in also removes references in the Windows library. Tested: Built with VS2005
* [svn-r14300] new feature: add a mention of multiple filters use in usagePedro Vicente Nunes2007-11-281-0/+4
| | | | tested: windows, linux
* [svn-r14299] new feature: add support for h5repack to handle several global ↵Pedro Vicente Nunes2007-11-2815-1963/+1759
| | | | | | | | filters usage is to repeat the -f option tested: windows, linux, solaris
* [svn-r14298] Purpose: Remove references of H5FD_stream.c from Windows ↵Scott Wegner2007-11-284-130/+0
| | | | | | | | | project files. Description: The stream file driver was recently removed from the HDF5-1.8 source. This checkin removes any left-over references in the Windows projects. Tested: VS2005
* [svn-r14297] Added the ability to locate szlib libraries in THG hosts.Albert Cheng2007-11-281-35/+102
| | | | | Tested: kagiso, linew. All worked. Smirom (configure works, make failed due to szlib itself),
* [svn-r14295] Description:Quincey Koziol2007-11-273-135/+284
| | | | | | | | | | | | | | | | | | | | | | - Eliminate some redundant calls to retrieve datatype sizes in chunk read/ write routines. - Change indexed storage "common" B-tree callback user data to avoid copying chunk offset and separate "downward" info from "upward" info. - Cache chunk info (nbytes/filter_mask/address) for last chunk accessed Tested on: FreeBSD/32 6.2 (duty) in debug mode FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe, in debug mode Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN, in production mode Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN, in production mode Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN, w/szip filter, in production mode Mac OS X/32 10.4.10 (amazon) in debug mode Linux/64-ia64 2.4 (tg-login3) w/parallel, w/FORTRAN, in production mode
* [svn-r14294] bug fix: in the filter percentage calculation Pedro Vicente Nunes2007-11-271-7/+14
| | | | tested: windows, linux, solaris
* [svn-r14293] bug fix: the printing of object name was not done when H5Ocopy ↵Pedro Vicente Nunes2007-11-271-3/+6
| | | | | | was used tested: windows, linux
* [svn-r14292] enhancement: clarify usage of the -s optionPedro Vicente Nunes2007-11-261-2/+2
| | | | tested: windows, linux
generated by cgit v0.12 at 2024-06-26 01:23:45 (GMT)