Changeset 1085

Show
Ignore:
Timestamp:
07/01/08 10:39:24 (2 months ago)
Author:
fcady
Message:

modified some text files

Files:

Legend:

Unmodified
Added
Removed
Modified
Copied
Moved
  • trunk/pyhdf/CHANGES

    r1054 r1085  
    1 # $Id: CHANGES,v 1.8 2008/06/23 
    2 # $Log: CHANGES,v $ 
    3 # Revision 1.8 Enthought 
    4 # pyhdf-0.7-4 
    5 # Ported to HDF4.2r1 
    6 # Uses the NumPy rather than the now obsolete Numeric 
    7  
    81# $Id: CHANGES,v 1.7 2005/07/14 01:36:40 gosselin_a Exp $ 
    92# $Log: CHANGES,v $ 
     
    3326List of changes to 'pyhdf' since the initial release. 
    3427 
    35 pyhdf 0.7-3 
    36   June 23 2008 
    37      - This release depends on the NumPy package, rather than the Numeric 
    38         package 
     28pyhdf-0.8 
     29--------- 
     30  July 1 2008 
     31    - pyhdf is now based on 'numpy'. Support for the now phased out 'Numeric' 
     32      package has been dropped. 
     33    - The swig interface file is now part of the distribution. 
     34    - Fixed bug in the 'setscale()' method of an SDim instance. 
     35    - Users working with a debian-based linux distro reported problems with 
     36      the HDF SDgetcompress() / SDsetcompress() compression-related functions. 
     37      The cause seems to be related to the use of an HDF library whose version 
     38      predates 4.2r1. Debian appears very slow moving to the most  
     39      recent HDF release. To avoid depriving debian users of the use of pyhdf, 
     40      macro 'NOCOMPRESS' can be set inside 'setup.py'. This will transform 
     41      functions SDsetcompress() and SDgetcompress() into no-ops which  
     42      will return -1 (HDF error code) if called. 
     43    - Examples of the use of the HDF compression functionnalities have been  
     44      completely reworked.  See directory 'examples/compress' for details. 
     45 
    3946 
    4047pyhdf 0.7-3 
  • trunk/pyhdf/INSTALL

    r1054 r1085  
    1 # $Id: INSTALL,v 1.4 2008/06/23        Enthought 
     1# $Id: INSTALL,v 1.4 2005/07/17 03:59:41 gosselin_a Exp $ 
    22# $Log: INSTALL,v $ 
    3 # Revision 1.4 2008/06/23 
    4 # pyhdf-0.7-4 
    5 # Ported to HDF4.2r1. 
    6 # NumPy replaces Numeric as underlying scientific package 
     3# Revision 1.4  2005/07/17 03:59:41  gosselin_a 
     4# Fixed typo in header line. 
    75# 
    8 # $Id: INSTALL,v 1.3 2005/07/14 01:36:41 gosselin_a Exp $ 
    9 # $Log: INSTALL,v $ 
    106# Revision 1.3  2005/07/14 01:36:41  gosselin_a 
    117# pyhdf-0.7-3 
     
    1410# All classes are now 'new-style' classes, deriving from 'object'. 
    1511# Update documentation. 
    16 # 
    17 # Revision 1.6                          Enthought 
    18 # pyhdf 0.7-4 
    1912# 
    2013# Revision 1.2  2004/08/02 15:16:38  gosselin 
     
    2518# 
    2619 
    27 Installing the pycdf python package. 
     20Installing the pyhdf python package. 
    2821 
    2922pyhdf has been successfully installed under python 2.2 and above on 
     
    3427  -AIX 4 
    3528 
    36 Please inform the author (rkern@enthought.com) of any 
     29Please inform the maintainer (rkern@enthought.com) of any 
    3730successfull installation on a different platform, and of 
    3831problems encountered. 
     
    7467    library_dirs = ["/usr/local/hdf-4.2r1/lib"], 
    7568 
     69 
     70  External libraries 
     71 
    7672  HDF4.2 no longer provides its own copies of the jpeg and z libraries. 
    7773  Those must be installed separately (on Linux, they should be part of 
    78   any standard distribution). The sz 2.0 library must be installed if the SZIP 
    79   compression method is to be used, and HDF must have been compiled with 
    80   SZIP support. Note that if you installed precompiled HDF binaries provided 
    81   by NCSA, you *must* install SZIP in a separate step (see NCSA hdf site for 
    82   details). In case your HDF library does not include SZIP support, 
    83   set the following inside file 'setup.py' (see comments inside this file) : 
     74  any standard distribution).  
    8475 
    85     extra_compile_args=["-DNOSZIP"] 
     76  The sz library (versions 2.0 or higher) must be installed if the SZIP  
     77  compression method is to be used with SDsetcompress(). HDF v4.2 must also  
     78  have been compiled with SZIP support. Precompiled HDF packages usually 
     79  lack this feature, as in Fedora linux distributions. Getting 
     80  an SZIP enabled HDF library may require compiling the library from source 
     81  with the "--with-szlib" configuration option. 
     82  Note that you *must* install SZIP in a separate step. For more details, 
     83  see NCSA hdf site at : http://hdf.ncsa.uiuc.edu/doc_resource/SZIP/ ; 
     84  source code and precompiled binaries are available at : 
     85  ftp://ftp.hdfgroup.org/lib-external/szip/ .  
     86 
     87  In case your HDF library was compiled without SZIP support, or you cannot 
     88  abide by the szip licensing terms, set the following inside file  
     89  'setup.py' (see comments inside this file) : 
     90 
     91    extra_compile_args = ["-DNOSZIP"] 
    8692    libraries = ["mfhdf", "df", "jpeg", "z"] 
     93 
     94  If you get error messages related to the SDgetcompress() / SDsetcompress() 
     95  functions, eg.: "undefined symbol: SDgetcompress" (this seems to affect 
     96  users of Debian based distros), try also setting the NOCOMPRESS macro : 
     97 
     98    extra_compile_args = ["-DNOSZIP", "-DNOCOMPRESS"] 
     99 
     100  With NOCOMPRESS set, SDgetcompress() and SDsetcompress() will be transformed 
     101  into no-ops which will immediately raise an exception, and will not be resolved 
     102  against the HDF library symbols. This may help you work with an HDF library 
     103  earlier than v4.2. 
     104 
     105 
     106  Swig-generated interface files 
     107 
     108    Interface files 'hdfext.py' and 'hdfext_wrap.c' (located under the 
     109    'pyhdf/' subdirectory) have been generated using the 'swig' tool. 
     110    Those two files should be usable as is on most environments. 
     111    It could happen however that, for a reason related to your environment, 
     112    your C compiler does not accept the '.c' file and provokes a compilation 
     113    error. If this is so, the interface files will need to be regenerated. 
     114    For this you will first need to have the swig tool installed 
     115    (see: http://www.swig.org). Then execute : 
     116 
     117      # cd pyhdf 
     118      # swig -python hdfext.i 
     119 
     120    swig should silently regenerate the two interface files, and the install should now 
     121    proceed correctly. 
    87122 
    88123 
     
    118153    (on linux: /usr/lib/pythonx.x). 
    119154 
     155   Warning messages about the implicit declaration of some functions can be produced. 
     156   Those are due to the way the swig tool generates the C code, and can be safely ignored. 
     157 
    1201586- Before experimenting with pyhdf, exit root privileges and leave the installation directory 
    121159   (the one where you untarred the product). Failing to do so may cause all sorts 
    122160   of mysterious problems. 
    123161 
     1627- As a first test, run the 'hdfstruct.py' script (found under directory 'examples/hdfstruct') 
     163   on one of your HDF4 files. The script should display the file structure. This is 
     164   a handy tool to have around when you want to explore the contents of any HDF4 file. 
     165 
    124166Enjoy! 
  • trunk/pyhdf/PKG-INFO

    r1055 r1085  
    11Metadata-Version: 1.0 
    2 Name: pyhdf  
    3 Version: 0.7-4 
     2Name: pyhdf 
     3Version: 0.8 
    44Summary: Python interface to the NCSA HDF4 library 
    55Home-page: www.sourceforge.net/projects/pysclint 
    6 Maintainer: Enthought Inc. 
    7 Maintainer email: enthought-dev@mail.enthought.com 
    8 Original Author: Andre Gosselin 
    9 Original Author-email: gosselina@dfo-mpo.gc.ca 
     6Author: Robert Kern 
     7Author-email: robert.kern@enthought.com 
    108License: public 
    119Description: The pyhdf package wraps the functionality 
     
    1311        framework. The SD (scientific dataset), VS 
    1412        (Vdata) and V (Vgroup) APIs are currently implemented. 
    15         Other APIs should be covered in the 
    16         near future. SD datasets are read/written 
    17         through NumPy arrays. netCDF files can also 
     13        SD datasets are read/written 
     14        through numpy arrays. netCDF files can also 
    1815        be read and modified with pyhdf. 
    19 Keywords: hdf, netcdf, NumPy, python 
     16        The function parse_odl is also provided to 
     17        deal specifically with data in the ODL 
     18        (Object Desdription Language) format. 
     19Keywords: hdf,netcdf,numpy,python,pyhdf,parse_odl 
    2020Platform: UNKNOWN