Building and installation instructions for Alpha Open VMS HDF5 1.8.12 release 18 September 2013 CONTENTS -------- 1. Obtaining HDF5 2. Building and testing HDF5 C, Fortran and C++ libraries, and utilities 3. Installing HDF5 libraries 4. Known problems 1.Obtaining HDF5 Please see INSTALL file in this directory for downloading instructions. 2. Building and testing HDF5 C, Fortran and C++ libraries and utilities. IMPORTANT: This version REQUIRES GNU ZLIB library to be installed on a system. You may download the source code from http://www.zlib.net/ The newest ZLIB 1.2.8 doesn't work properly on OpenVMS. But ZLIB 1.2.5 works fine. - Use tar command to untar HDF5 source ball tar -xvf .tar - Change default directory to [..vms] - Specify top HDF5 source directory in the build.com file where the "hdf5top" is defined. - Edit make.com: specify path to ZLIB library if necessary, modify compilation flags Do NOT REMOVE /define=H5_VMS and /standard=strict_ansi qualifiers. - Run build.com command file @build We recommend using batch queue to build and test this distribution. The command file runs make.com and builds C library, tests, and utilities Fortran library and tests C++ library and tests It also runs test scripts to test C, Fortran and C++ libraries, and h5dump, h5diff, h5ls, and h5repack utilities. 3. Installing HDF5 libraries Modify install.com file to specify installation directory (hdf5top) and run @install to install HDF5 C, Fortran, and C++ libraries, and utilities. The script will create the directories shown below in the installation directory: [.HDF5]... [.HDF5.INCLUDE] [.HDF5.LIB] [.HDF5.BIN] [.HDF5.EXAMPLES] [.HDF5.EXAMPLES.C] [.HDF5.EXAMPLES.F90] [.HDF5.EXAMPLES.CXX] [.HDF5.EXAMPLES.*] directories contain make.com and check.com files to build examples against the installed libraries and to verify the installation. 4. Known problems - There is no support for szip compression filter. - Two soft conversion functions (H5T__conv_i_f and H5T__conv_f_i) have bugs. They convert data between floating-point numbers and integers. But the library's default is hard conversion. The user should avoid explicitly enabling soft conversion between floating-point numbers and integers. For help contact help@hdfgroup.org