diff --git a/.github/workflows/linkchecker.yml b/.github/workflows/linkchecker.yml new file mode 100644 index 00000000000..681ceaf4dbf --- /dev/null +++ b/.github/workflows/linkchecker.yml @@ -0,0 +1,47 @@ +name: linkchecker + +on: + workflow_dispatch: + push: + pull_request: + branches: [ develop ] + +concurrency: + group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} + cancel-in-progress: true + +permissions: + contents: read + +jobs: + build: + runs-on: ubuntu-latest + steps: + - name: get + uses: actions/checkout@v4 + + - name: install + run: | + sudo apt update -y + sudo apt install -y libunwind-dev + sudo apt install -y graphviz + sudo apt install -y --no-install-recommends doxygen + sudo apt install -y cmake + sudo apt install -y texlive-full + sudo apt install -y texlive-fonts-extra + sudo apt install -y texlive-latex-extra + pip3 install linkchecker + + - name: build + run: | + mkdir build + cd build + cmake -DHDF5_BUILD_DOC:BOOL=ON .. + make doxygen + + - name: check + run: | + cd build/hdf5lib_docs/html + linkchecker --check-extern ./index.html + continue-on-error: true + diff --git a/.github/workflows/release-files.yml b/.github/workflows/release-files.yml index a0e1e367b07..721f10fa102 100644 --- a/.github/workflows/release-files.yml +++ b/.github/workflows/release-files.yml @@ -221,7 +221,7 @@ jobs: - name: PreRelease tag id: create_prerelease if: ${{ (inputs.use_environ == 'snapshots') }} - uses: softprops/action-gh-release@e7a8f85e1c67a31e6ed99a94b41bd0b71bbee6b8 # v2.0.9 + uses: softprops/action-gh-release@01570a1f39cb168c169c802c3bceb9e93fb10974 # v2.1.0 with: tag_name: "${{ inputs.use_tag }}" prerelease: true @@ -249,7 +249,7 @@ jobs: - name: Release tag id: create_release if: ${{ (inputs.use_environ == 'release') }} - uses: softprops/action-gh-release@e7a8f85e1c67a31e6ed99a94b41bd0b71bbee6b8 # v2.0.9 + uses: softprops/action-gh-release@01570a1f39cb168c169c802c3bceb9e93fb10974 # v2.1.0 with: tag_name: "${{ inputs.use_tag }}" prerelease: false diff --git a/.github/workflows/scorecard.yml b/.github/workflows/scorecard.yml index a90240dd273..ed87159e5c7 100644 --- a/.github/workflows/scorecard.yml +++ b/.github/workflows/scorecard.yml @@ -67,6 +67,6 @@ jobs: # Upload the results to GitHub's code scanning dashboard. - name: "Upload to code-scanning" - uses: github/codeql-action/upload-sarif@662472033e021d55d94146f66f6058822b0b39fd # v3.27.0 + uses: github/codeql-action/upload-sarif@f09c1c0a94de965c15400f5634aa42fac8fb8f88 # v3.27.5 with: sarif_file: results.sarif diff --git a/config/cmake/hdf5-config.cmake.in b/config/cmake/hdf5-config.cmake.in index 02286b701a6..6908db269cd 100644 --- a/config/cmake/hdf5-config.cmake.in +++ b/config/cmake/hdf5-config.cmake.in @@ -88,6 +88,7 @@ if (${HDF5_PACKAGE_NAME}_ENABLE_PARALLEL) set (${HDF5_PACKAGE_NAME}_PARALLEL_FILTERED_WRITES @PARALLEL_FILTERED_WRITES@) set (${HDF5_PACKAGE_NAME}_LARGE_PARALLEL_IO @LARGE_PARALLEL_IO@) + enable_language(C) # for MPI::MPI_C find_package(MPI QUIET REQUIRED) endif () diff --git a/doxygen/dox/Overview.dox b/doxygen/dox/Overview.dox index 6caffd147aa..bbdb6c6466c 100644 --- a/doxygen/dox/Overview.dox +++ b/doxygen/dox/Overview.dox @@ -21,7 +21,7 @@ documents cover a mix of tasks, concepts, and reference, to help a specific audience succeed. \par Offline reading - You can download it as an archive for offline reading. + You can download it as an archive for offline reading. \par ToDo List There is plenty of unfinished business. diff --git a/doxygen/dox/TechnicalNotes.dox b/doxygen/dox/TechnicalNotes.dox index b0f1c8f4cbc..8bea54bfb25 100644 --- a/doxygen/dox/TechnicalNotes.dox +++ b/doxygen/dox/TechnicalNotes.dox @@ -1063,7 +1063,161 @@ been used at the lowest level to process the request. /** \page APPDBG Debugging HDF5 Applications -\htmlinclude DebuggingHDF5Applications.html +\section sec_adddbg_intro Introduction +The HDF5 library contains a number of debugging features to make programmers' lives +easier including the ability to print detailed error messages, check invariant +conditions, display timings and other statistics. + +\subsection subsec_adddbg_intro_err Error Messages +Error messages are normally displayed automatically on the standard error stream and +include a stack trace of the library including file names, line numbers, and function +names. The application has complete control over how error messages are displayed and +can disable the display on a permanent or temporary basis. Refer to the documentation + for the H5E error handling package. + +\subsection subsec_adddbg_intro_invar Invariant Conditions +Unless NDEBUG is defined during compiling, the library will include code to verify that +invariant conditions have the expected values. When a problem is detected the library will +display the file and line number within the library and the invariant condition that +failed. A core dump may be generated for post mortem debugging. The code to perform these +checks can be included on a per-package bases. + +\subsection subsec_adddbg_intro_stats Timings and Statistics +The library can be configured to accumulate certain statistics about things like cache +performance, datatype conversion, data space conversion, and data filters. The code is +included on a per-package basis and enabled at runtime by an environment variable. + +\subsection subsec_adddbg_intro_trace API Tracing +All API calls made by an application can be displayed and include formal argument names +and actual values and the function return value. This code is also conditionally included +at compile time and enabled at runtime. + +The statistics and tracing can be displayed on any output stream (including streams opened by +the shell) with output from different packages even going to different streams. + +\section sec_adddbg_msg Error Messages +By default any API function that fails will print an error stack to the standard error stream. +\code +HDF5-DIAG: Error detected in thread 0. Back trace follows. + #000: H5F.c line 1245 in H5Fopen(): unable to open file + major(04): File interface + minor(10): Unable to open file + #001: H5F.c line 846 in H5F_open(): file does not exist + major(04): File interface + minor(10): Unable to open file +\endcode +The error handling package (H5E) is described elsewhere. + +\section sec_adddbg_invars Invariant Conditions +To include checks for invariant conditions the library should be configured +with --disable-production, the default for versions before 1.2. The library +designers have made every attempt to handle error conditions gracefully but +an invariant condition assertion may fail in certain cases. The output from +a failure usually looks something like this: +\code +Assertion failed: H5.c:123: i + +Name + +Default + +Description + + + +aNoAttributes + + +acYesMeta data cache + + +bYesB-Trees + + +dYesDatasets + + +eYesError handling + + +fYesFiles + + +gYesGroups + + +hgYesGlobal heap + + +hlNoLocal heaps + + +iYesInterface abstraction + + +mfNoFile memory management + + +mmYesLibrary memory management + + +oNoObject headers and messages + + +pYesProperty lists + + +sYesData spaces + + +tYesDatatypes + + +vYesVectors + + +zYesRaw data filters + + + +In addition to including the code at compile time the application must enable each package at +runtime. This is done by listing the package names in the HDF5_DEBUG environment variable. That +variable may also contain file descriptor numbers (the default is '2') which control the output +for all following packages up to the next file number. The word 'all' refers to all packages. Any +word my be preceded by a minus sign to turn debugging off for the package. + +\subsection subsec_adddbg_stats_sample Sample debug specifications + + + + + + + + + + + + + +
all +This causes debugging output from all packages to be sent to the standard error stream. +
all -t -s +Debugging output for all packages except datatypes and data spaces will appear on the standard error stream. +
-all ac 255 t,s +This disables all debugging even if the default was to debug something, then output +from the meta data cache is send to the standard error stream and output from data types +and spaces is sent to file descriptor 255 which should be redirected by the shell. +
+The components of the HDF5_DEBUG value may be separated by any non-lowercase letter. */ diff --git a/doxygen/dox/UsersGuide.dox b/doxygen/dox/UsersGuide.dox index 03151473d2a..14481a10f69 100644 --- a/doxygen/dox/UsersGuide.dox +++ b/doxygen/dox/UsersGuide.dox @@ -335,6 +335,23 @@ \ref sec_map +\section sec_hl HDF5 High Level APIs +\li \ref H5DO_UG + +\li \ref H5DS_UG + +\li \ref H5IM_UG +\li \ref H5LT_UG +\li \ref H5TB_UG +\li \ref H5PT_UG + \ref sec_cltools \li \ref H5TOOL_CP_UG \li \ref H5TOOL_DF_UG diff --git a/doxygen/dox/VOLConnGuide.dox b/doxygen/dox/VOLConnGuide.dox index 547d6eb76cb..00dd18e82f7 100644 --- a/doxygen/dox/VOLConnGuide.dox +++ b/doxygen/dox/VOLConnGuide.dox @@ -191,7 +191,8 @@ are for connector testing and should not be found in the wild. Values of 512 to connectors. As is the case with HDF5 filters, The HDF Group can assign you an official VOL connector value. Please -contact help@hdfgroup.org for help with this. We currently do not register connector names, though the +contact help@hdfgroup.org for help with this. We currently do not +register connector names, though the name you've chosen will appear on the registered VOL connectors page. As noted above, registered VOL connectors will be listed at: diff --git a/doxygen/examples/DebuggingHDF5Applications.html b/doxygen/examples/DebuggingHDF5Applications.html deleted file mode 100644 index 495a87c6676..00000000000 --- a/doxygen/examples/DebuggingHDF5Applications.html +++ /dev/null @@ -1,394 +0,0 @@ - - - Debugging HDF5 Applications - - -
-

Introduction

- -

The HDF5 library contains a number of debugging features to - make programmers' lives easier including the ability to print - detailed error messages, check invariant conditions, display - timings and other statistics, and trace API function calls and - return values. - -

-
Error Messages -
Error messages are normally displayed automatically on the - standard error stream and include a stack trace of the library - including file names, line numbers, and function names. The - application has complete control over how error messages are - displayed and can disable the display on a permanent or - temporary basis. Refer to the documentation for the H5E error - handling package. - -

-
Invariant Conditions -
Unless NDEBUG is defined during compiling, the - library will include code to verify that invariant conditions - have the expected values. When a problem is detected the - library will display the file and line number within the - library and the invariant condition that failed. A core dump - may be generated for post mortem debugging. The code to - perform these checks can be included on a per-package bases. - -

-
Timings and Statistics -
The library can be configured to accumulate certain - statistics about things like cache performance, datatype - conversion, data space conversion, and data filters. The code - is included on a per-package basis and enabled at runtime by - an environment variable. - -

-
API Tracing -
All API calls made by an application can be displayed and - include formal argument names and actual values and the - function return value. This code is also conditionally - included at compile time and enabled at runtime. -
- -

The statistics and tracing can be displayed on any output - stream (including streams opened by the shell) with output from - different packages even going to different streams. - -

Error Messages

- -

By default any API function that fails will print an error - stack to the standard error stream. - -

-

- - - - -
- 


-HDF5-DIAG: Error detected in thread 0.  Back trace follows.
-  #000: H5F.c line 1245 in H5Fopen(): unable to open file
-    major(04): File interface
-    minor(10): Unable to open file
-  #001: H5F.c line 846 in H5F_open(): file does not exist
-    major(04): File interface
-    minor(10): Unable to open file
-
-
-
- -

The error handling package (H5E) is described - elsewhere. - -

Invariant Conditions

- -

To include checks for invariant conditions the library should - be configured with --disable-production, the - default for versions before 1.2. The library designers have made - every attempt to handle error conditions gracefully but an - invariant condition assertion may fail in certain cases. The - output from a failure usually looks something like this: - -

-

- - - - -
- 


-Assertion failed: H5.c:123: i<NELMTS(H5_debug_g)
-IOT Trap, core dumped.
-
-
-
- -

Timings and Statistics

- -

Code to accumulate statistics is included at compile time by - using the --enable-debug configure switch. The - switch can be followed by an equal sign and a comma-separated - list of package names or else a default list is used. - -

-

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
NameDefaultDescription
aNoAttributes
acYesMeta data cache
bYesB-Trees
dYesDatasets
eYesError handling
fYesFiles
gYesGroups
hgYesGlobal heap
hlNoLocal heaps
iYesInterface abstraction
mfNoFile memory management
mmYesLibrary memory management
oNoObject headers and messages
pYesProperty lists
sYesData spaces
tYesDatatypes
vYesVectors
zYesRaw data filters
-
- -

In addition to including the code at compile time the - application must enable each package at runtime. This is done - by listing the package names in the HDF5_DEBUG - environment variable. That variable may also contain file - descriptor numbers (the default is `2') which control the output - for all following packages up to the next file number. The - word all refers to all packages. Any word my be - preceded by a minus sign to turn debugging off for the package. - -

-

- - - - - - - - - - - - - - -
Sample debug specifications
allThis causes debugging output from all packages to be - sent to the standard error stream.
all -t -sDebugging output for all packages except datatypes - and data spaces will appear on the standard error - stream.
-all ac 255 t,sThis disables all debugging even if the default was to - debug something, then output from the meta data cache is - send to the standard error stream and output from data - types and spaces is sent to file descriptor 255 which - should be redirected by the shell.
-
- -

The components of the HDF5_DEBUG value may be - separated by any non-lowercase letter. - -

API Tracing

- -

The HDF5 library can trace API calls by printing the - function name, the argument names and their values, and the - return value. Some people like to see lots of output during - program execution instead of using a good symbolic debugger, and - this feature is intended for their consumption. For example, - the output from h5ls foo after turning on tracing, - includes: - -

-

- - - - -
- 
-H5Tcopy(type=184549388) = 184549419 (type);
-H5Tcopy(type=184549392) = 184549424 (type);
-H5Tlock(type=184549424) = SUCCEED;
-H5Tcopy(type=184549393) = 184549425 (type);
-H5Tlock(type=184549425) = SUCCEED;
-H5Fopen(filename="foo", flags=0, access=H5P_DEFAULT) = FAIL;
-HDF5-DIAG: Error detected in thread 0.  Back trace follows.
-  #000: H5F.c line 1245 in H5Fopen(): unable to open file
-    major(04): File interface
-    minor(10): Unable to open file
-  #001: H5F.c line 846 in H5F_open(): file does not exist
-    major(04): File interface
-    minor(10): Unable to open file
-
-
-
- -

The code that performs the tracing must be included in the - library by specifying the --enable-trace - configuration switch (the default for versions before 1.2). Then - the word trace must appear in the value of the - HDF5_DEBUG variable. The output will appear on the - last file descriptor before the word trace or two - (standard error) by default. - -

-

- - - - - - - -
To display the trace on the standard error stream: - 
$ env HDF5_DEBUG=trace a.out
-
-
To send the trace to a file: - 
$ env HDF5_DEBUG="55 trace" a.out 55>trace-output
-
-
-
- -

Performance

- -

If the library was not configured for tracing then there is no - unnecessary overhead since all tracing code is excluded. - However, if tracing is enabled but not used there is a small - penalty. First, code size is larger because of extra - statically-declared character strings used to store argument - types and names and extra auto variable pointer in each - function. Also, execution is slower because each function sets - and tests a local variable and each API function calls the - H5_trace() function. - -

If tracing is enabled and turned on then the penalties from the - previous paragraph apply plus the time required to format each - line of tracing information. There is also an extra call to - H5_trace() for each API function to print the return value. - -

Safety

- -

The tracing mechanism is invoked for each API function before - arguments are checked for validity. If bad arguments are passed - to an API function it could result in a segmentation fault. - However, the tracing output is line-buffered so all previous - output will appear. - -

Completeness

- -

There are two API functions that don't participate in - tracing. They are H5Eprint() and - H5Eprint_cb() because their participation would - mess up output during automatic error reporting. - -

On the other hand, a number of API functions are called during - library initialization and they print tracing information. - -

Implementation

- -

For those interested in the implementation here is a - description. Each API function should have a call to one of the - H5TRACE() macros immediately after the - FUNC_ENTER() macro. The first argument is the - return type encoded as a string. The second argument is the - types of all the function arguments encoded as a string. The - remaining arguments are the function arguments. This macro was - designed to be as terse and unobtrousive as possible. - -

In order to keep the H5TRACE() calls synchronized - with the source code we've written a perl script which gets - called automatically just before Makefile dependencies are - calculated for the file. However, this only works when one is - using GNU make. To reinstrument the tracing explicitly, invoke - the trace program from the hdf5 bin directory with - the names of the source files that need to be updated. If any - file needs to be modified then a backup is created by appending - a tilde to the file name. - -

-

- - - - - -
Explicit Instrumentation
- 
-$ ../bin/trace *.c
-H5E.c: in function `H5Ewalk_cb':
-H5E.c:336: warning: trace info was not inserted
-
-
-
- -

Note: The warning message is the result of a comment of the - form /*NO TRACE*/ somewhere in the function - body. Tracing information will not be updated or inserted if - such a comment exists. - -

Error messages have the same format as a compiler so that they - can be parsed from program development environments like - Emacs. Any function which generates an error will not be - modified.

-
- diff --git a/doxygen/examples/H5DS_Spec.pdf b/doxygen/examples/H5DS_Spec.pdf deleted file mode 100644 index 813f4ded3e1..00000000000 Binary files a/doxygen/examples/H5DS_Spec.pdf and /dev/null differ diff --git a/doxygen/img/H5DS_fig1.png b/doxygen/img/H5DS_fig1.png new file mode 100644 index 00000000000..f47619eb6db Binary files /dev/null and b/doxygen/img/H5DS_fig1.png differ diff --git a/doxygen/img/H5DS_fig2.png b/doxygen/img/H5DS_fig2.png new file mode 100644 index 00000000000..6eace5a9d05 Binary files /dev/null and b/doxygen/img/H5DS_fig2.png differ diff --git a/doxygen/img/H5DS_fig3.png b/doxygen/img/H5DS_fig3.png new file mode 100644 index 00000000000..d0771268fdc Binary files /dev/null and b/doxygen/img/H5DS_fig3.png differ diff --git a/doxygen/img/H5DS_fig4.png b/doxygen/img/H5DS_fig4.png new file mode 100644 index 00000000000..6e1de71c73f Binary files /dev/null and b/doxygen/img/H5DS_fig4.png differ diff --git a/doxygen/img/H5DS_fig5.png b/doxygen/img/H5DS_fig5.png new file mode 100644 index 00000000000..3d1e7ecf241 Binary files /dev/null and b/doxygen/img/H5DS_fig5.png differ diff --git a/hl/src/H5DSpublic.h b/hl/src/H5DSpublic.h index a926035a4c8..f38f6fef0ba 100644 --- a/hl/src/H5DSpublic.h +++ b/hl/src/H5DSpublic.h @@ -31,6 +31,725 @@ extern "C" { #endif /** \page H5DS_UG HDF5 High Level Dimension Scales + * + * \section sec_dim_scales_stand HDF5 Standard for Dimension Scales + * Dimension scales are stored as datasets, with additional metadata indicating that they are to + * be treated as dimension scales. Each dimension scale has an optional name. There is no requirement + * as to where dimension scales should be stored in the file. Dimension Scale names are not required + * to be unique within a file. (The name of a dimension scale does not have to be the same as the HDF5 + * path name for the dataset representing the scale.) + * + * Datasets are linked to dimension scales. Each dimension of a Dataset may optionally have one or + * more associated Dimension Scales, as well as a label for the dimension. A Dimension Scale can be + * shared by two or more dimensions, including dimensions in the same or different dataset. + * Relationships between dataset dimensions and their corresponding dimension scales are not be directly + * maintained or enforced by the HDF5 library. For instance, a dimension scale would not be automatically + * deleted when all datasets that refer to it are deleted. + * + * Functions for creating and using Dimension Scales are implemented as high level functions, see \ref H5DS. + * + * A frequently requested feature is for Dimension Scales to be represented as functions, rather than a + * stored array of precomputed values. To meet this requirement, it is recommended that the dataset + * model be expanded in the future to allow datasets to be represented by a generating function. + * + * \section sec_dim_scales_concept Conceptual model + * Our study of dimension scale use cases has revealed an enormous variety of ways that dimension + * scales can be used. We recognize the importance of having a model that will be easy to understand + * and use for the vast majority of applications. It is our sense that those applications will need + * either no scale, a single 1-D array of floats or integers, or a simple function that provides a + * scale and offset. + * + * At the same time, we want to place as few restrictions as possible on other uses of dimension + * scales. For instance, we don’t want to require dimension scales to be 1-D arrays, or to allow + * only one scale per dimension. + * + * So our goal is to provide a model that serves the needs of two communities. We want to keep + * the dimension scale model conceptually simple for the majority of applications, but also to + * place as few restrictions as possible how dimension scales are interpreted and used. With + * this approach, it becomes the responsibility of applications to make sure that dimension scales + * satisfy the constraints of the model that they are assuming, such as constraints on the size + * of dimension scales and valid range of indices. + * + * \subsection subsec_dim_scales_concept_defs Definitions + * Dimension Scales are implemented as an extension of these objects. In the HDF5 Abstract Data + * Model, a Dataset has a Dataspace, which defines a multi dimensional array of elements. Conceptually, + * a Dataspace has N dimension objects, which define the current and maximum size of the array in + * that dimension. + * + * It is important to emphasize that the Dataspace of a Dataset has no intrinsic meaning except + * to define the layout in computer storage. Dimension Scales may be used to store application + * specific labels to the positions in the stored data array, i.e., to add application specific + * meaning to the dimensions of the dataspace. + * + * A Dimension Scale is an object associated with one dimension of a Dataspace. The meaning of + * the association is left to applications. The values of the Dimension Scale are set by the + * application to reflect semantics of the data, for example, to associate coordinates of a + * reference system with positions on the dimension. + * + * In general, these associations define a mapping between values of a dimension index and + * values of the Dimension Scale dataset. A simple case is where the Dimension Scale s is + * a (one dimensional) sequence of labels for the dimension ix of Dataset d. In this case, + * Dimension Scale is an array indexed by the same index as in the dimension of the Dataspace. + * For example, for the Dimension Scale s, associated with dimension ix, the ith position of + * ix is associated with the value s[i], so s[i] is taken as a label for ix[i]. + * + * \subsection subsec_dim_scales_concept_rel Entity Relationship Diagrams + * Figure 1 shows UML to illustrate the relationship between a Dimension and a Dimension + * Scale object. Conceptually, each Dimension of a Dataspace may have zero or more Dimension + * Scales associated with it. In turn, a Dimension Scale object may be associated with zero + * or more Dimensions (in zero or more Dataspaces). + * + * Figure 2 illustrates the abstract model for a Dimension Scale object. A Dimension Scale + * is represented as a sub-class of a Dataset: a Dimension Scale has all the properties of + * a Dataset, with some specializations. A Dimension Scale dataset has an attribute “CLASS” + * with the value “DIMENSION_SCALE”. (This is analogous to the Table, Image, and Palette + * objects.) The Dimension Scale dataset has other attributes, including an optional + * NAME and references to any associated Dataset, as discussed below. + * + * When the Dimension Scale is associated with a dimension of a Dataset, the association + * is represented by attributes of the two datasets. In the Dataset, the DIMENSION_LIST + * is an array of object references to scales (Dimension Scale Datasets) (Figure 1), and + * in the Dimension Scale Dataset the REFERENCE_LIST is an array of object references + * to Datasets (Figure 2). + * + * + * + * + *
+ * \image html H5DS_fig1.png "Figure 1. The relationship between a Dimension and a Dimension Scale." + *
+ * + * + * + * + * + *
+ * \image html H5DS_fig2.png "Figure 2. The definition of a Dimension Scale and its attributes." + *
+ * + * \subsection subsec_dim_scales_concept_types What types of scales should be implemented? + * There seems to be good agreement that the model should accommodate scales that consist of a + * stored 1-D list of values, certain simple functions, and “no scale.” This specification also + * includes scales that are higher dimensional arrays, as well. + * + * The four types of scales are: + * \li No scale. Frequently no scale is needed, so it should not be required. + * In some of these cases, an axis label may still be needed, and should be available. + * In this case, the Datase * t defines a Dimension Scale label for a dimension, with no + * Dimension Scale dataset. + * \li 1-D array. Both fixed length and extendable arrays should be available. + * The size of the Dimension Scale is not required by HDF5 to conform to the size of the + * corresponding dimension(s), so that the number of scale values could be less than, + * equal to, or greater than the corresponding dimension size. + * \li Simple function. At a minimum, a linear scale of the form A + Bx + * should be available. This will be discussed in a future proposal. + * \li Higher dimensional arrays. This specification allows Dimension Scales + * to have any number of dimensions. + * + * A number of use cases have been proposed in which more than one scale is needed for a given dimension. + * This specification places no restrictions on the number of scales that can be associated with a dimension, + * nor on the number or identities of Dimensions that may share the same Dimension Scale. + * + * There are use cases for storing many types of data in a scale, including, + * but not limited to integers, floats, and strings. Therefore, this specification places no restrictions + * on the datatypes of scale values: a Dimension Scale can have any HDF5 Datatype. The + * interpretation of dimension scale values is left to applications. + * + * \subsection subsec_dim_scales_concept_limits Limitations of this Specification + * One-to-many mapping. When there are fewer values in a dimension scale than in the corresponding + * dimension, it is useful to have a mapping between the two. For example, mappings are used by + * HDF-EOS to map geolocation information to dimension, in order that, for example, every second + * point may have geolocation. On the other hand, the way that mappings are defined can be very + * idiosyncratic, and it would seem to be challenging to provide a mapping model that satisfied a + * large number of cases. These mappings are not included in the model specified here. + * + * Visibility and Integrity. Since Dimension Scales are a specialization of a Dataset, it is + * “visible” and accessible as a regular Dataset through the HDF5 API. This means that an application + * program could alter the values of or delete a Dimension Scale object or required attributes without + * regard to any of the semantics defined in this document. + * + * One advantage it that the implementation requires no changes to the base library, which reduces + * the complexity of the code and the risk of side-effects. The implementation builds on existing + * functions, which should improve the quality and reliability of the code. Also, this approach leaves + * most of the semantics of dimension scales to applications and communities, who can use the + * specification in any way they need. + * + * An important disadvantage is that the core HDF5 library will not manage the semantics of + * Dimension Scales. In particular, applications or other software must implement: + * \li Naming – the HDF5 library will impose no rules on the names of Dimension Scales + * \li Consistency of references – e.g., if a Dataset (Dimension Scale) is deleted (e.g., + * with #H5Ldelete, any Dimension Scales (Datasets) that it refers to (refer to it) + * will not be updated by the HDF5 library. + * \li Consistency of extents – the HDF5 library will not assure that a Dimension and + * associated Dimension Scale have the same extent (number of elements), nor that shared + * objects are consistent with each other. As in the case of delete, if a Dimension or + * Dimension Scale is extended (e.g., H5S…), any associated objects will not be + * automatically extended. + * + * These are briefly summarized here. + * \li Naming and Name Spaces. There are many potential schemes for naming dimensions, each + * suited for different uses. This specification does not impose any specific approach, + * so it may be used by different applications. However, the lack of restrictions has + * disadvantages as well.
For some purposes, it will be important to iterate through + * all the Dimension Scale objects in a file. The iterate operation is difficult to + * implement with the design specified here. This will be left to other software. For example, + * the HDF-EOS library has its own mechanism for managing a set of dimensions, and the + * netCDF4 library will implement this if it needs to. + * \li Automatically extending dataset dimensions. When a dimension of a dataset is + * extended, should the library automatically extend the corresponding dimension scale, or + * should this be left to the application? Since a dimension scale can be shared among many + * datasets, this raises a number of issues that are difficult to address in a general way. + * For instance, which dimension scale should be extended when only one dataset is extended, + * and what values are to be added? We have seen no compelling reason to implement an automatic + * extension of dimension scales when dataset dimensions are extended, so we suggest letting + * applications be responsible for this operation. + * \li Automatically deleting dimension scales. Should a dimension scale be deleted + * when all datasets that use it have been deleted? This is another case where different applications + * might have different requirements, so a general policy would be difficult to devise. Furthermore, + * enforcing a deletion policy, even a simple one, adds complexity to the library, and could also + * affect performance. Deletion policies seem best left to applications. + * + * Section \ref sec_dim_scales_api presents an API and programming model that implements some of these + * features. However, applications + * may ignore or bypass these APIs, to write or read the attributes directly. + * + * \section sec_dim_scales_spec The HDF5 Dimension Scale Specification + * + * \subsection subsec_dim_scales_spec_summary Brief Summary + * A Dimension Scale is stored as an HDF5 Dataset. + * \li A Dimension Scale is an object that is associated with a dimension of a Dataset. + * \li A Dimension Scale can have at most one name. + * \li A Dimension Scale may be associated with zero, one, or many different dimensions in any number of + * Datasets. + * \li Unless otherwise specified, a Dimension Scale inherits the properties of an HDF5 Dataset. + * \li There are no restrictions on the size, shape, or datatype of a Dimension Scale. + * + * A Dimension Scale can be associated with a dimension of an HDF5 dataset + * \li A dimension of a Dataset may have zero, one, or more Dimension Scales associated with it. + * \li Each scale is identified by an index value. + * + * A dimension may have a label without a scale, and may have a scale with no label. + * \li The label need not be the same as the name of any associated Dimension Scales. + * + * The implementation has two parts: + * \li A storage profile + * \li An API and programming model + * + * \subsection subsec_dim_scales_spec_store + * This section specifies the storage profile for Dimension Scale objects and the association between + * Dimensions and Dimension Scales. + * + * This profile is compatible with an earlier netcdf prototype and the HDF4 to HDF5 Mapping. This + * profile is also compatible with the netCDF4 proposal. This profile may be used to augment the + * HDF-EOS5 profile. + * + * See Appendix 2 for a discussion of how to store converted HDF4 objects. See Appendix 3 for a + * discussion of netCDF4 issues. See Appendix 4 for a discussion of HDF-EOS5. + * + * \subsubsection subsubsec_dim_scales_spec_store_dset + * A Dimension Scale dataset is stored as an HDF5 dataset. Table 1 summarizes the stored data, i.e., + * the values of the scale. There are no restrictions on the dataspace or datatype, or storage properties of + * the dataset. + * + * The scale may have any HDF5 datatype, and does not have to be the same as the datatype of + * the Dataset(s) that use the scale. E.g., an integer dataset might have dimension scales + * that are string or float values. + * + * The dataspace of the scale can be any rank and shape. A scale is not limited to one dimension, + * and is not restricted by the size of any dimension(s) associated with it. When a dimension is + * associated with a one dimensional scale, the scale may be a different size from the dimension. + * In this case, it is up to the application to interpret or resolve the difference. When a + * dimension is associated with a scale with a rank higher than 1, the interpretation of the + * association is up to the application. + * + * The Dimension Scale dataset can use any storage properties (including fill values, filters, + * and storage layout), not limited by the properties of any datasets that refer to it. When + * the Dimension Scale is extendible, it must be chunked. + * + * Table 2 defines the required and optional attributes of the Dimension Scale Dataset. The + * attribute REFERENCE_LIST is a list of (dataset, index) pairs. Each pair represents an + * association defined by ‘attach_scale’. These pairs are stored as an array of compound data. + * Table 3 defines this datatype. + * + * The Dimension Scale Dataset has an attribute called SUB_CLASS. This string is intended to + * be used to document particular specializations of this profile, e.g., a Dimension Scale + * created by netCDF4. + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 1. The properties of the Dimension Scale dataset
FieldDatatypeDataspaceStorage PropertiesNotes
<data>AnyAnyAnyThese are the values of the Dimension Scale.
+ * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 2. Standard Attributes for a stored Dimension Scale dataset.
Attribute NameDatatype and DimensionsValueRequired / OptionalNotes
CLASS#H5T_STRING length = 16“DIMENSION_SCALE”RequiredThis attribute distinguishes the dataset as a Dimension scale object.
This is set by + * #H5DSset_scale
NAME#H5T_STRING length = <user defined><user defined>
The name does not have to be the same as the HDF5 path name for the dataset. + * The name does not have to be related to any labels. Several Dimension Scales may have the same name.		Optional, (Maximum of 1)The user defined label of the Dimension Scale.
This is set by #H5DSset_label
REFERENCE_LISTArray of Dataset Reference Type (Compound Datatype), variable length.[ {dataset1, ind1 }, …] [,…] ….]Optional, required when scale is attachedSee Table 3. This is set by #H5DSattach_scale.
SUB_CLASS#H5T_STRING length = <profile defined>“HDF4_DIMENSION”,
“NC4_DIMENSION”,		Optional, defined by other profilesThis is used to indicate a specific profile was used.
<Other attributes>OptionalFor example, UNITS.
+ * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 3. Dataset Reference Type. + * This is a pair, <dataset_ref, index>. This is created when the Dimension Scale is attached to a + * Dataset.
FieldDatatypeValueNotes
DATASETObject Reference.Pointer to a Dataset that refers to the scaleSet by #H5DSattach_scale.
Removed by #H5DSdetach_scale.
INDEX#H5T_NATIVE_INTIndex of the dimension the dataset pointed to by DATASETSet by #H5DSattach_scale.
Removed by #H5DSdetach_scale.
+ * + * \subsubsection subsubsec_dim_scales_spec_store_attr Attributes of a Dataset with a Dimension Scale + * A Dataset may have zero or more Dimension Scales associated with its dataspace. When present, + * these associations are represented by two attributes of the Dataset. Table 4 defines these attributes. + * + * The DIMENSION_LIST is a two dimensional array with one row for each dimension of the Dataset, and + * a variable number of entries in each row, one for each associated scale. This is stored as a one + * dimensional array, with the HDF5 Datatype variable length array of object references. + * + * When a dimension has more than one scale, the order of the scales in the DIMENSION_LIST attribute + * is not defined. A given Dimension Scale should appear in the list only once. + * (I.E., the DIMENSION_LIST is a “set”.) + * + * When a scale is shared by more than one dimension (of one or more Dataset), the order of + * the records in REFERENCE_LIST is not defined. The Dataset and Dimension should appear in the list only + * once. + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 4. Standard Attributes of a Dataset with associated Dimension Scale.
Attribute NameDatatype and DimensionsValueRequired / OptionalNotes
DIMENSION_LISTThe HDF5 datatype is ARRAY of Variable Length #H5T_STD_REF_OBJ with rank of the dataspace.[[{object__ref1, object__ref2, … object__refn}, …] […] ..]Optional, required if scales are attachedSet by #H5DSattach_scale.
Entries removed by #H5DSdetach_scale.
DIMENSION_LABELLISTThe HDF5 datatype is ARRAY of #H5T_STRING with rank of the dataspace.[ <Label1>, <Label2>, …, <Label3>] Optional, required for scales with a labelSet by #H5DSset_label.
+ * + * \subsection subsec_dim_scales_spec_lab Dimension Scale Names and Labels + * Dimension scales are often referred to by name, so Dimension Scales may have names. + * Since some applications do not wish to apply names to dimension scales, Dimension Scale + * names be optional. In addition, some applications will have a name but no associated data + * values for a dimension (i.e., just a label). To support this, each dimension may have a + * label, which may be but need not be the same as the name of an associated Dimension Scale. + * + * Dimension Scale Name. Associated with the Dimension Scale object. A Dimension Scale may + * have no name, or one name. + * + * Dimension Label. A optional label associated with a dimension of a Dataset. + * + * How is a name represented? Three options seem reasonable: + * \li Last link in the pathname. The h4toh5 mapping uses this approach [6], but there could be more + * than one path to a dataset, leading to ambiguities. This could be overcome by enforcing conventions. + * \li Attribute. This exposes this information at the top level, making it accessible to any viewer that + * can see attributes. It also makes it easy for applications to change the name, which could be + * dangerous, or valuable. + * \li Header message. This approach makes the name a little less available at the top level, + * but firmly pushes the concept into the base format and library. Since it also requires + * applications to change the name through a function call, it leaves open the possibility + * that the form of the name could be altered later without requiring application codes + * to change. On the other hand, if we treat names this way, it means that the “name” + * attribute is being treated differently from the “class” attribute, which could be confusing. + * + * Dimension Scale names are stored in attributes of the Dimension Scale or the Dataset that refers + * to a Dimension Scale. + * + * Should dimension scale names be unique among dimension scales within a file? + * We have seen a number of cases in which applications need more than one dimension scale with + * the same name. We have also seen applications where the opposite is true: dimension scale + * names are assumed to be unique within a file. This specification leaves it to applications + * to enforce a policy of uniqueness when they need it. + * + * Can a dimension have a label, without having an associated scale? Some + * applications may wish to name dimensions without having an associated scale. Therefore, + * a dataset may have a label for a dimension without having an associated Dimension Scale dataset. + * + * Can a dimension have a scale, without having an associated label? Some + * applications may wish to assign a dimension scale with no label. Therefore, a dataset + * may have one or more Dimension Scales for a dimension without having an associated label. + * + * Anonymous Dimensions. It is possible to have a Dimension Scale dataset + * with no name, and associate it with a dimension of a dataset with no label. This case + * associates an array of data values to the dimension, but no identifier. + * + * A dimension with a label and a name. A dimension of a dataset can be + * associated with a Dimension Scale that has a name, and assigned a label. In this case, + * the association has two “names”, the label and the dimension scale name. It is up to + * applications to interpret these names. + * + * Table 5 summarizes the six possible combinations of label and name. + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 5. Labels and scales of a dimension.
No scaleScale with no nameScale with name
No labelDimension has no label or scale (default)Dimension has an anonymous scaleDimension has scale, the scale is called “name”
LabelDimension has labelDimension has scale with a label.Dimension has scale with both a label and name. A shared dimension has one name, + * but may have several labels
+ * + * \subsection subsec_dim_scales_spec_shared Shared Dimension Scales + * Given the design described above, datasets can share dimension scales. The following + * additional capabilities would seem to be useful. + * \li When a dimension scale is deleted, remove the reference to the dimension scale in + * all datasets that refer to it. + * \li Determine how many datasets are attached to a given dimension scale + * \li Determine what datasets are attached to a given dimension scale + * + * These capabilities can be provided in several ways: + * \li Back pointers. If every dimension scale contained a list of back pointers to all + * datasets that referenced it, then it would be relatively easy to open all of these + * datasets and remove the references, as well as to answer questions #2 and #3. + * This would require the library to update the back pointer list every time a link + * was made. + * \li Alternatively, such lists could be maintained in a separate table. Such a table + * could contain all information about a number of dimension scales, which might provide + * a convenient way for applications to gain information about a set of dimension scales. + * For instance, this table might correspond to the coordinate variable definitions in a netCDF file. + * \li If no such list were available, an HDF5 function could be available to search all datasets + * to determine which ones referenced a given dimension scale. This would be straightforward, + * but in some cases could be very time consuming. + * + * This specification defines attributes that maintain back pointers, which enable + * these kinds of cross referencing. Other software, such as NetCDF4, may well need a global table to + * track a set of dimensions. Such a table can be done in addition to the attributes defined here. + * + * \subsection subsec_dim_scales_spec_ex Example + * This section presents an example to illustrate the data structures defined above. + * + * Figure 3 shows a Dataset with a four dimensional Dataspace. The file also contains six Dimension + * Scale datasets. The Dimension Scale datasets are HDF5 objects, with path names such as “/DS1”. + * + * Figure 4 illustrates the use of dimension scales in this example. Each Dimension Scale Dataset + * has an optional NAME. For example, “/DS3” has been assigned the name “Scale3”. + * + * The dimensions of dataset D have been assigned zero or more scales and labels. Dimension 0 + * has two scales, Dimension 1 has one scale, and so on. Dimension 2 has no scale associated with it. + * + * Some of the dimensions have labels as well. Note that dimension 2 has a label but no scale, and + * dimension 3 has scales but no label. + * + * Some of the Dimension Scales are shared. Dimension Scale DS1 is referenced by dimension 0 of D + * and by another unspecified dataset. Dimension Scale DS3 is referenced by dimension 1 and 3 of Dataset D. + * + * These relationships are represented in the file by attributes of the Dataset D and the Dimension + * Scale Datasets. Figure 5 shows the values that are stored for the DIMENSION_LIST attribute of + * Dataset D. This + * + * + * + * + *
+ * \image html UML_Attribute.jpg "The UML model for an HDF5 attribute" + *
+ * attribute is a one-dimensional array with the HDF5 datatype variable length + * #H5T_STD_REF_OBJ. Each row of the array is zero or more object references for Dimension Scale datasets. + * + * Table 6 shows the DIMENSION_LABELLIST for Dataset D. This is a one dimensional array with some empty + * values. + * + * Each of the Dimension Scale Datasets has a name and other attributes. The references are + * represented in the REFERENCE_LIST attributes. Table 7 – Table 10 show the values for these + * tables. Note that Dimension Scale DS4 and DS6 have no references to them in this diagram. + * + * The tables are stored as attributes of the Dimension Scale Dataset and the Datasets that + * refer to scales. Essentially, the association between a dimension of a Dataset and a Dimension + * Scale is represented by “pointers” (i.e., HDF5 Object References) in both of the associated + * objects. Since there can be multiple associations, there can be multiple pointers stored at + * each object, representing the endpoints of the associations. These will be stored in tables, + * i.e., as an attribute with an array of values. + * + * When dimension scales are attached or detached, the tables in the Dataset and the Dimension + * Scale must be updated. The arrays in the attributes can grow, and items can be deleted. + * + * The associations are identified by the object reference and dimension which is stored in a + * back pointer and returned from an API. The detach function needs to be careful how it deletes + * an item from the table, because the entries at both ends of the association must be updated + * at the same time. + * + * + * + * + * + *
+ * \image html H5DS_fig3.png "Figure 3. Example dataset and scales." + *
+ * + * + * + * + * + *
+ * \image html H5DS_fig4.png "Figure 4. Example labels, names, and attached scales." + *
+ * + * + * + * + * + *
+ * \image html H5DS_fig5.png "Figure 5. The table of dimension references, stored as an attribute of the + * Dataset." + *
+ * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 6. The table of dimension labels.
Dataset DimensionLabel
0“LX”
1“LZ”
2“LQ”
3“”
+ * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 7. The reference list for DS1.
ReferenceDataset Reference Record
0{Object reference to Dataset D, 0}
1{Object reference to other Dataset, ?}
+ * + * + * + * + * + * + * + * + * + * + *
Table 8. Reference list for DS2
ReferenceDataset Reference Record
0{Object reference to Dataset D, 0}
+ * + * + * + * + * + * + * + * + * + * + * + * + * + * + *
Table 9. Reference list for DS3
ReferenceDataset Reference Record
0{Object reference to Dataset D, 1}
1{Object reference to Dataset D, 3}
+ * + * + * + * + * + * + * + * + * + * + *
Table 10. Reference List for DS5
ReferenceDataset Reference Record
0{Object reference to Dataset D, 3}
+ * + * \section sec_dim_scales_api Programming Model and API + * + * \subsection subsec_dim_scales_api_model Programming Model + * Dimension Scales are HDF5 Datasets, so they may be created and accesses through any + * HDF5 API for datasets [10]. The HDF5 Dimension Scale API implements the specification + * defined in this document. The operations include: + * \li Convert dataset to scale (D) – convert dataset D to a dimension scale. D may be + * specified by an id or by a path name. + * \li Attach scale (D, S, i) – attach dimension scale S to the ith dimension of D. D + * and S may be specified by an id or path name + * \li Detach scale (D, i, scale) – detach scale from the ith dimension of D. + * \li Iterate through scales of (D, i) – get each scale attached. + * \li Get the number of scales of (D, i) + * \li Get the ith scale of D + * \li Set/Get name (S) – set/get the name about dimension scale S. + * + * The API also defines operations for dimension labels: + * \li Set/Get label (D, i) – set/get the label for ith dimension of D. + * + * \subsubsection subsubsec_dim_scales_api_model_create Create new Dimension Scale with Initial Values + * 1. Create dataset with for the Dimension Scale with H5Dcreate and other standard HDF5 calls. + * 2. Initialize the values of the Dimension Scale with H5Dwrite and other calls. + * 3. Convert the dataset to a Dimension Scale with H5DSmake_scale. + * 4. Close the Dimension Scale when finished with H5Dclose. + * + * \subsubsection subsubsec_dim_scales_api_model_attach Attach Dimension Scale to Dataset + * 1. Create or open the Dataset, D, with H5Dopen, etc. + * 2. Create or open the Dimension Scale dataset, S, with H5Dopen or as above. + * 3. Attach the Dimension Scale S to dimension j of Dataset D with H5DSattach_scale + * 4. When finished, close the Dimension Scale and Dataset with H5Dclose. + * + * \subsubsection subsubsec_dim_scales_api_model_read Read Dimension Scale values + * 1. Open the Dataset D, with H5Dopen + * 2. Get the number of dimensions of D with H5Dget_space. + * 3. Iterate through the scales of dimension i, locate the target scale, S (e.g., by its name). + * 4. Get the datatype, dataspace, etc. of S with H5Dget_space, H5Dget_type, H5Sget_ndims, etc. + * 5. Read the values of S into memory with H5Dread, e.g. into dscalebuff. + * 6. When finished, close the S and other objects with H5Dclose etc. + * 7. When finished, close the Dataset D with H5Dclose. + * + * \subsubsection subsubsec_dim_scales_api_model_write Write or Update Dimension Scale values + * 1. Open the Dimension Scale Dataset S with H5open + * 2. Get the datatype, dataspace, etc. of S with H5Dget_space, H5Dget_type, H5Sget_ndims, etc. + * 3. If needed, read the values of S into memory with H5Dread. Note, may read selected values using a + * selection. + * 4. Write updated values to S with H5Dwrite. Note, may write selected values using a selection. + * 5. When finished, close S and other objects with H5Dclose etc. + * + * \subsubsection subsubsec_dim_scales_api_model_label Create a label for a dimension + * 1. Open the Dataset D with H5open + * 2. Add write a label for dimension i of D, with H5DSset_label. + * 3. When finished, close the Dimension Scale Dataset and other objects with H5Dclose etc. + * + * \subsubsection subsubsec_dim_scales_api_model_extend Extending a Dimension with a Dimension Scale attached + * When an extendible Dataset has Dimension Scales, it is necessary to coordinate when the dimensions change + * size. + * 1. Open the Dataset to be extended, with H5Dopen. + * 2. Extend the dimension(s) with H5Dextend + * 3. Iterate through the scales of each extended dimension. For each scale + * a. Extend the scale to the new size of the dimension with H5Dextend + * b. Write new values to the extended scale with H5Dwrite, etc. + * c. Close the Dimension Scale Dataset with H5Dclose if necessary. + * 4. When finished, close the Dataset with H5Dclose + * + * \subsubsection subsubsec_dim_scales_api_model_detach Detach Dimension Scale from Dataset + * The detach operation removes an association between a dimension and a scale. It does not delete the + * Dimension Scale Dataset. + * 1. Open the Dataset, D, with H5Dopen. + * 2. Iterate through the scales of dimension i, locate the target scale, S (e.g., by its name). + * 3. Detach the Dimension Scale S to dimension j of Dataset D with H5DSdetach_scale + * 4. When finished, close the Dimension Scale and Dataset with H5Dclose. + * + * \subsubsection subsubsec_dim_scales_api_model_del Delete a Dimension Scale Dataset + * When it is necessary to delete a Dimension Scale Dataset, it is necessary to detach it from all dataset. + * This section outlines the necessary steps. + * 1. Open the Dimension Scale to be deleted. + * 2. Read the REFERENCE_LIST attribute into memory with H5Aread etc. + * 3. For each entry in the list: + * a. Dereference the dataset reference + * b. Detach the scale with H5DSdetach_scale + * c. Close the dataset reference + * 4. Delete the Dimension Scale Dataset + * + * \subsubsection subsubsec_dim_scales_api_model_clean Clean up Dimension Scales when deleting a Dataset + * When it is necessary to delete a dataset that has scales attached, it is necessary to delete all the + * scales before deleting the dataset. Here is a sketch of the steps. + * 1. Open the Dataset to be deleted, with H5Dopen. + * 2. Iterate through the scales of each dimension of D + * 3. For each scale, detach the Dimension Scale S from dimension j of Dataset D with H5DSdetach_scale + * 4. Delete the Dataset, with H5Gunlink. + * + * \subsection subsec_dim_scales_api_func Programming API: H5DS + * @see H5DS Reference Manual + * * @todo Under Construction */ diff --git a/release_docs/RELEASE.txt b/release_docs/RELEASE.txt index 130fa59482c..47fd4637ef9 100644 --- a/release_docs/RELEASE.txt +++ b/release_docs/RELEASE.txt @@ -561,6 +561,12 @@ Bug Fixes since HDF5-2.0.0 release Testing ------- + - Added skipping of a few parallel tests for OpenMPI 5.0.5 + + An issue in OpenMPI 5.0.5 causes a few parallel HDF5 tests + (mpiodup, props, fapl_preserve) to fail. These tests are + now skipped for that release of OpenMPI. The issue has + been fixed in the 5.0.6 release of OpenMPI. Platforms Tested diff --git a/src/H5Iint.c b/src/H5Iint.c index aeb8ed25935..5be8dcd38a8 100644 --- a/src/H5Iint.c +++ b/src/H5Iint.c @@ -1495,7 +1495,12 @@ H5I__iterate_cb(void *_item, void H5_ATTR_UNUSED *_key, void *_udata) object = H5I__unwrap(info->u.object, type); /* Invoke callback function */ - cb_ret_val = (*udata->user_func)((void *)object, info->id, udata->user_udata); + /* Prepare & restore library for user callback */ + H5_BEFORE_USER_CB_NOERR(H5_ITER_ERROR) + { + cb_ret_val = (*udata->user_func)((void *)object, info->id, udata->user_udata); + } + H5_AFTER_USER_CB_NOERR(NULL) /* Set the return value based on the callback's return value */ if (cb_ret_val > 0) diff --git a/src/H5Ppublic.h b/src/H5Ppublic.h index 34e4d76181f..be6ea9fe3d2 100644 --- a/src/H5Ppublic.h +++ b/src/H5Ppublic.h @@ -3913,8 +3913,7 @@ H5_DLL herr_t H5Pget_mdc_image_config(hid_t plist_id, H5AC_cache_image_config_t * access property list, and H5Fget_mdc_logging_status() will return * the current state of the logging flags. * - * The log format is described in the - * Metadata Cache Logging document. + * The log format is described in the \ref_mdc_logging document. * * \since 1.10.0 */ diff --git a/src/H5TSint.c b/src/H5TSint.c index cb52df51170..b803891de50 100644 --- a/src/H5TSint.c +++ b/src/H5TSint.c @@ -333,7 +333,8 @@ H5TS_api_lock(void) done: FUNC_LEAVE_NOAPI_NAMECHECK_ONLY(ret_value) } /* end H5TS_api_lock() */ -#else /* H5_HAVE_CONCURRENCY */ +#else +#ifdef H5_HAVE_CONCURRENCY herr_t H5TS_api_lock(unsigned *dlftt) { @@ -362,6 +363,9 @@ H5TS_api_lock(unsigned *dlftt) done: FUNC_LEAVE_NOAPI_NAMECHECK_ONLY(ret_value) } /* end H5TS_api_lock() */ +#else +#error "Unknown multithreading mode" +#endif #endif /*-------------------------------------------------------------------------- diff --git a/src/H5module.h b/src/H5module.h index 35429b57a15..a5a448af6bf 100644 --- a/src/H5module.h +++ b/src/H5module.h @@ -593,7 +593,7 @@ * * * * *
- * \image html Dmodel_fig14_c.gif " Another HDF5 file structure with groups and datasets" + * \image html Dmodel_fig14_d.gif " Another HDF5 file structure with groups and datasets" *
diff --git a/testpar/testphdf5.c b/testpar/testphdf5.c index c511ee387be..3253776dfb5 100644 --- a/testpar/testphdf5.c +++ b/testpar/testphdf5.c @@ -346,7 +346,19 @@ main(int argc, char **argv) test_params.count = 0; /* Tests are generally arranged from least to most complexity... */ + + /* Skip this test for OpenMPI version 5.0.5 since an issue was + * introduced in that release which causes this test to fail. + * The issue was fixed for the 5.0.6 release. + */ +#if defined(OPEN_MPI) && defined(OMPI_MAJOR_VERSION) && (OMPI_MAJOR_VERSION == 5) && \ + defined(OMPI_MINOR_VERSION) && (OMPI_MINOR_VERSION == 0) && defined(OMPI_RELEASE_VERSION) && \ + (OMPI_RELEASE_VERSION == 5) + AddTest("-mpiodup", test_fapl_mpio_dup, NULL, NULL, NULL, 0, "fapl_mpio duplicate"); +#else AddTest("mpiodup", test_fapl_mpio_dup, NULL, NULL, NULL, 0, "fapl_mpio duplicate"); +#endif + AddTest("getdxplmpio", test_get_dxpl_mpio, NULL, NULL, &test_params, sizeof(test_params), "dxpl_mpio get"); @@ -357,8 +369,19 @@ main(int argc, char **argv) AddTest("page_buffer", test_page_buffer_access, NULL, NULL, &test_params, sizeof(test_params), "page buffer usage in parallel"); + /* Skip this test for OpenMPI version 5.0.5 since an issue was + * introduced in that release which causes this test to fail. + * The issue was fixed for the 5.0.6 release. + */ +#if defined(OPEN_MPI) && defined(OMPI_MAJOR_VERSION) && (OMPI_MAJOR_VERSION == 5) && \ + defined(OMPI_MINOR_VERSION) && (OMPI_MINOR_VERSION == 0) && defined(OMPI_RELEASE_VERSION) && \ + (OMPI_RELEASE_VERSION == 5) + AddTest("-props", test_file_properties, NULL, NULL, &test_params, sizeof(test_params), + "Coll Metadata file property settings"); +#else AddTest("props", test_file_properties, NULL, NULL, &test_params, sizeof(test_params), "Coll Metadata file property settings"); +#endif AddTest("delete", test_delete, NULL, NULL, &test_params, sizeof(test_params), "MPI-IO VFD file delete"); @@ -367,8 +390,20 @@ main(int argc, char **argv) AddTest("evictparassert", test_evict_on_close_parallel_unsupp, NULL, NULL, &test_params, sizeof(test_params), "Evict on close in parallel failure"); + + /* Skip this test for OpenMPI version 5.0.5 since an issue was + * introduced in that release which causes this test to fail. + * The issue was fixed for the 5.0.6 release. + */ +#if defined(OPEN_MPI) && defined(OMPI_MAJOR_VERSION) && (OMPI_MAJOR_VERSION == 5) && \ + defined(OMPI_MINOR_VERSION) && (OMPI_MINOR_VERSION == 0) && defined(OMPI_RELEASE_VERSION) && \ + (OMPI_RELEASE_VERSION == 5) + AddTest("-fapl_preserve", test_fapl_preserve_hints, NULL, NULL, &test_params, sizeof(test_params), + "preserve MPI I/O hints after fapl closed"); +#else AddTest("fapl_preserve", test_fapl_preserve_hints, NULL, NULL, &test_params, sizeof(test_params), "preserve MPI I/O hints after fapl closed"); +#endif AddTest("idsetw", dataset_writeInd, NULL, NULL, &test_params, sizeof(test_params), "dataset independent write");