- Getting compression type for JPEG chunked images is not working yet.
- For a JPEG image, GRgetcompress only returns the compression type, not
the compression information (i.e, quantity and force_baseline). This
information is not currently retrievable.
- When HDF 4.1r5 came out, it was not tested with gcc 3.2 on Linux.
However, it should build with a few changes. Here are some
problems that you may encounter:
- Undefined reference to `fabs'
- In function `strerror': : `sys_errlist' is deprecated; use `strerror' or `strerror_r' instead
- When running "make test", the compile of cdftest fails with the error "undefined reference to `errno'".
These problems can be resolved by changes to the ./HDF4.1r5/config/mh-linux file. Edit this file, make the necessary changes, then build the software. (Be sure to run "make distclean" if using the same source code as before.)
For the fabs problem:
In mh-linux, search on "fabs". You will see the following line:#LIBSX = -lm
Remove the "#" (in other words, uncomment the line):LIBSX=-lm
For the strerror problem:
In mh-linux, search for STRERROR. You will find:# for Sunos and Linux CPPFLAGS_HDF = $(FMPOOL_FLAGS) -DNDEBUG -DHDF -DNO_STRERROR
Remove the -DNO_STRERROR option from CPPFLAGS_HDF.Further above this line it gives more information:
# Additional flags for preproccesor. Some no longer used because # library now requires ANSI compilier. # # for no function prototypes add -DNO_HAVE_PROTOTYPES # for no strerror() add -DNO_STRERROR
For the undefined reference to 'errno':
In mh-linux, search for CFLAGS. Add -DPC to the CFLAGS variable.To build the software:
./configure -v --prefix=<installpath for HDF> make >& comp.out make test >& test.out make install
If you already built the software once, you should run "make distclean" first. - HDF 4.1r5 was not tested with the Intel compiler on Windows.
When building with the Intel compiler all tests pass except for
the N-bit test (see the function, SDsetnbitdataset). We have
fixed this problem and will be providing the pre-compiled binaries when
we can (9/16/02).
Following are instructions on using the Intel compiler with MSVC++:
1. Invoke MVSC++ 2. Go to Tools -->Select Compiler; "Selection Tool" window will come up. Check the box with Intel C++ compiler and click OK. Intel C++ compiler now will be used when application is built with MVSC++. 3. To use Fortran Intel compiler do the following: GO to Tools --> Customize... The "Customize" window will come up. Click on "Add-ins and Macros Files" tab and check "Intel Fortran Compiler Build Tool" box
- - Last modified:June 25th 2007
