Revisions of python-pysndfile

Dominique Leuenberger's avatar Dominique Leuenberger (dimstar_suse) accepted request 818481 from Steve Kowalik's avatar Steve Kowalik (StevenK) (revision 5)
- Update to 1.4.3:
  * changed sndio functions to all use PySndfile as context manager. This fixes the problem that the sndfile
    remains open when an error occurs which may in turn lead to inconsistencies if the sndfile is tried to be rewritten
    in an exception handler.
  * fixed PySndfile.read_frames method to properly handle reading frames in parts (previous fix was incomplete)
  * extended supported commands to change compression level when writing flac and ogg files
  * fixed PySndfile.read_frames and sndio.read method to properly handle reading frames from the middle of a file
  * Extended PySndfile class:
     * support use as context manager
     * added support for wve, ogg, MPC2000 sampler and RF64 wav files
     * added support for forcing to return 2D arrays even for mono files
     * added method to close the file and release all resources.
     * support reading more frames than present in the file using the fill_value for all values positioned after the end of the file
  * (no changes in functionality)
  * added documentation to distributed files
  * added missing licence file to distribution
  * thanks @toddrme2178 for patches.
  * removed cython (a build requirement) from requirements.txt
  * avoid cython warning and fix language_level in the .pyx source code
  * add and support pre-release tags in the version number
  * use hashlib to calculate the README checksum.
  * fixed support for use with python 2.7 that was broken since 1.3.4
  * fixed potential but undesired build dependency of pandoc
  * added link to explanation for using pysndfile under windows
  * fixed pandoc problem that does produce non ASCII chars in rst output.
  * fixed two copy paste bug introduced in 1.3.4
  1.3.4 did in fact not work at all :-(
  * added a check target to the makefile that performs a complete built/install/test cycle
  to avoid problems as in 1.3.4
  * added support for automatic installation of requirements
  * remove precompiled cython source file and rely on pip requirements to provide cython
    so that cython compilation will always be possible.
  * added experimental support for installation on win32 (thanks
    to Svein Seldal for the contributions). 
  * use expanduser for replacing ~ in filenames
  * adapted cython source code to avoid all compiler warnings due to deprecated numpy api 
  * removed use of ez_setup.py that is no longer required.
  * fixed missing command C\_SFC\_SET\_SCALE\_INT\_FLOAT\_WRITE (thanks
    to Svein Seldal for the bug report and fix)
  * better documentation of sf\_string-io in sndio.read and sndio.write
  * limit size of strings to be written such that the written file can
    always be read back with libsndfile 1.0.28 (which imposes different
    constraints for different formats)
  * better error handling when number of channels exceeds channel limit
    imposed by libsndfile.
  * sndio module now exposes the dicts: fileformat\_name\_to\_id
    and fileformat\_id\_to\_name 
  * extended sndio.read with force_2d argument that can be used to
    force the returned data array to always have 2 dimensions even for
    mono files.
- Drop extra source files of COPYING.txt and COPYING.LESSER.txt, it is now
  included.
- Add patch fix-test-imports.patch to fix the test suite.
Dominique Leuenberger's avatar Dominique Leuenberger (dimstar_suse) accepted request 628717 from Tomáš Chvátal's avatar Tomáš Chvátal (scarabeus_iv) (revision 2)
- Remove the test condition, run them always

- Update to version 1.3.2
  * No obvious changelog
Dominique Leuenberger's avatar Dominique Leuenberger (dimstar_suse) accepted request 605515 from Todd R's avatar Todd R (TheBlackCat) (revision 1)
reading/writing soundfiles in python
Displaying all 7 revisions
openSUSE Build Service is sponsored by