news

大名鼎鼎的CE下播放软件,TCPPMP的源代码!!!2410下可以流畅的解QVGA的H264,MPEG4等格式.

Summary of changes between libmikmod 3.1.11 and libmikmod 3.2.0 ():
==========================================================================

THANKS

- Axel "awe" Wefers <[e-mail]>,
  for creating the CoreAudio driver, the new
  AIFF driver and updating the Carbon driver.
- Frank Loemker, for sending me his collection of patches
  for libmikmod.
- Anders F Bj鰎klund <[e-mail]>, for updates, fixes and tests on
  Mac Platforms.

NEW FEATURES

- MacOS X CoreAudio driver
- Carbon support
- Loader for ASYLUM Music Format v1.0 (.amf) modules,
  as used in crusader games.
- New function 'Player_QueryVoices' added, which is used to return
  dynamic channel information about the module player. (Mikmod 3.2.2 uses
  it for it's dynamic panels).
- Mixing with floating point output. The MacOSX CoreAudio driver uses it. 
  Thanks goes to Anders F Bj鰎klund for implementing this.
- AIFF disk writer driver.  

BUGFIXES

- The sun driver now compiles on solaris (Thanks to Nicolas Gros d'Aillon
  for sending me Jonathan Perkin's fix)


Summary of changes between libmikmod 3.1.10 and libmikmod 3.1.11 (Thiers):
==========================================================================

libmikmod 3.1.11 was released on 01/21/2004.

This is a small maintenance release. I know there are not many changes,
but since it has been a very long time since the last one, and meanwhile
libmikmod changed maintaner, I think it is justified.

THANKS

- Ingo Saitz, the maintainer of the mikmod packages for debian, for
  bugfixes.
- Frank Loemker, for the lcc-win32 stuff and fixes for the direct sound
  driver.

BUGFIXES

- Playback problems in MODs of length >128. This fixes the file beatwave.mod. Thanks
  to Emmanuel Coirier for pointing this out.
- Applied debian patches:
  - Fix for broken volume fadeouts of IT instruments
  - Fixed configure test for pthread
  - devfs support for Linux OSS

- Makefile for lcc-win32. To compile libmikmod with lcc-win32 a rather new
    version is needed. I currently use version 3.7 compiled on Jan 20 2002.
    A version half a year old did not work.
    To compile
      - check that you have the lcc bin directory in the PATH
        (and no other compiler bin dirs),
      - change to libmikmod-3.1.11/win32,
      - check Makefile.lcc for any needed changes, and
      - use make -f Makefile.lcc.

  - Different fixes for the direct sound driver:
    - Pausing did only work for half of the buffer.
    - Restarting the player did not work as the output thread was removed
      in DS_PlayStop().
    - Removed busy waiting for end of thread.
    - Sound output is started immediately on the first call to DS_Update()
      to be consistent with other non threaded drivers.
    - Different small clean ups.

  - In MikMod_Reset() the raw output driver driver did not use the
    filename set via the cmdline.


Summary of changes between libmikmod 3.1.9 and libmikmod 3.1.10 (Arlanc):
=========================================================================

libmikmod 3.1.10 was released on 01/24/2002.

This version is yet another bugfix release. I have more contributions coming,
especially a native MacOS X driver, that will not be present in 3.1.10 but
will be in 3.1.11.


THANKS

- Most of the bugs mentioned below have been squished by ``UFO'',
  Holger Jakob, Thomas Neumann, Mark Page and Juan Linietsky.


BUGFIXES

- Systematic coredumps on library initialization and calls to
  MikMod_InfoDriver have been eliminated. Sorry, and thanks to Shlomi
  Fish for the best fix.

- The fix for spurious effect memory for portamentos in MOD files shouldn't
  have been applied to the ``porta to note'' (effect 3) effect. Fixed in this
  release. In the process, the aforementioned fix was perfected, as it
  broke a few modules by accident.

- IT Panbrello effect with random wave did not work.

- IT Tempo slide (T0x, T1x) was not updated accurately, and the effect memory
  was not implemented. The only tempo slide command which worked accurately
  was T10, which is a no-op (-:

- IT effect SC0 (cut note with no delay) is now correctly interpreted as SC1.

- XM effect 6 (porta + volume slide) didn't have memory.

- ProTracker effect F (change speed or bpm) should now be correctly handled
  when its value is 32, depending on the module type (some formats will use
  it as a speed value, others as a bpm value).
  
- XM playback stopped when reaching pattern 255 in the order list.

- Corrupted envelopes in XM instruments are detected and some repair is
  attempted.

- Some valid AMF, IMF and IT modules refused to load.

- Fine volume slide effects in volume column of IT modules, when using the
  effect memory (i.e. A0 and B0) were misinterpreted as AF and BF, thus
  causing unwanted large slides.

- Portamento to note effect in volume column of IT modules was not rendered
  correctly.

- S3M files marked as created by ScreamTracker will honour the limit of 64000
  bytes (not 64KB!) for sample length and sample offset effect position.

- Loops in instrument envelopes were processed one tick too fast.

- In-song message in FAR modules were not displayed correctly.

- ProTracker effect E9 (Retrig) was not played correctly.


NEW FEATURES

- Modules which are not panning-aware (not using any panning effects and
  with the tracker's default panning) use a recentered panning, so as to
  sound better with headphones. Suggested by Urpo Lankinen.

- Retrigger and panning effects are now recognized in FAR modules.

- All arpeggio effects are now played in OKT modules.

- FLT8 modules (8 channel StarTrekker) are now recognized and played.


PLATFORM SPECIFIC

- The Windows DirectSound driver has been improved and will use less
  resources during playback.

- The native audio driver for NetBSD and OpenBSD should work better with
  recent versions of these systems, as well as with old sound cards such
  as the SoundBlaster Pro.

- Mac OS X is supported, although no native sound driver is available for
  it yet, thanks to Juha Kaunisto.

- Dynamically loading drivers is now supported on HP-PA systems running
  under HP-UX, as well on systems using the a.out binary file format.



Summary of changes between libmikmod 3.1.8 and libmikmod 3.1.9 (Vorey):
=======================================================================

libmikmod 3.1.9 was released on 02/14/2000.

This version is a bugfix release, and was primarily considered as a public
patch for 3.1.8 (namely, version 3.1.8-a). However, due to the very important
bug fix in XM playback, among other things, it's worth being 3.1.9...


THANKS

- Many thanks to ``UFO'', Thomas Neumann, James Kehl and Vito Caputo for their
  help, code and feedback.
  
- Special thanks to Gerd Rausch, for writing both the Linux sam9407 driver and
  the associated Mikmod driver. Nice work !
  
- Thanks to Bastien Nocera for letting me do some tests under Linux/PPC.


BUGFIXES

- Envelopes weren't loaded correctly for IMF and XM modules. In the average
  case, the first envelope point was correct, the remaining half points were
  bogus, and the other half completely random. This fix is definitely worth
  the upgrade.
 
- In MOD files, effect 0xc (set channel volume) with effect parameter 0xff was
  misplayed as a note cut.
  
- In MOD files too, effects 0x1, 0x2 and 0x3 (portamentos) with effect
  parameter 0x00 were using the command memory, but these commands do not have
  effect memory for MOD files.
  
- The Wuzamod check for STM modules wasn't present in 3.1.8, despite what the
  release notes said )-:
 
- Enabling the envelope (with an effect) for a sample with no associated
  envelope would cause immediate coredumps. Now fixed.
  
- Panning was ignored in GDM modules.

- Sample rate was incorrect for STM and STX modules.

- Arpeggio memory (for IT modules) was ignored.

- Tempo slide for IT modules was not processed correctly.

- Toggling surround sound didn't have effect until the driver was reset.


NEW FEATURES

- New loader for Amiga Oktalyzer (.OKT) modules.

- UNI loader recognizes recent APlayer modules (.APUN format level 0x104).

- Due to popular request, the function Player_LoadTitleFP() has been added to
  the library.
  
- Under Unix systems, make targets "check" and "uninstall" have been added.
  However, due to a bug in automake, "make uninstall" will only work correctly
  with the next release of automake (or the current CVS tree).
  
- New output driver for sam9407-based soundcards under Linux, featuring
  hardware playback.


PLATFORM SPECIFIC

- Support for the latest Linux Ultra driver development versions (2.99) has
  been added. Also, a few bugs in this driver have been fixed.

- The DirectSound driver (win32) shipped with the previous release didn't
  compile. It should now - at least things are better !

- The OSS driver from 3.1.8 didn't work very well with some versions of the
  Linux kernel, as well as under Linux/PPC. The problems should be gone now.
  Also, the OSS driver now handles multi-soundcard configurations.



Summary of changes between libmikmod 3.1.7 and libmikmod 3.1.8 (Tence):
=======================================================================

libmikmod 3.1.8 was released on 10/25/1999.


THANKS

- Many thanks (in no particular order) to Frank Loemker, Bjornar Henden, Joshua
  Shagam, Vito Caputo, Thomas Neumann, Zach Metzinger, and Kevin Langman for
  their help, code and feedback.

- Special thanks to Andrew Zabolotny for his huge work on the OS/2 drivers.
  The rest of your code is not lost, it will go into 3.2...


BUGFIXES (more details can be found in file ``PROBLEMS'')

- IT global volume slide effect was played too fast.

- IT effect S6x (pattern delay for x ticks) was misunderstood as SEx (pattern
  delay for x rows).

- Finetune effect was not handled correctly for some XM modules.

- Software mixer incorrectly produced extra clicks in interpolated mode (fix by
  Frank Loemker).

- Under Unix systems, the piped output driver (drv_pipe), as well as the output
  to file drivers (drv_raw, drv_wav) had security issues if the libmikmod
  application is setuid root.

- Trying to read data from a file, with the data not located at offset 0,
  didn't work correctly since 3.1.5.

- Almost any text file beginning with "if" was considered as a valid module by
  Player_LoadTitle(), yet hopefully rejected when trying to load it.