Released version 0.7.13 (see the .13) with some networking code updates - they are still unfinished but pretty good already.
Yep we have another friday's update which is known to usually have some problems - as I no longer have access to RedHat box so I can't really check how the RedHat compiler works with avifile (as for some mysterious reason the friend whos box I've been using for this purpose is now running Debian as well :)) so if anyone wants to give me free account where I could test the compilability from time to time it might be good.
So what's new in this release: as usual I can't forget to say :) - tons of bugfixes - but there are also some quite important ones - finaly there is a method to retrieve CodecInfo list in some regular way. If you are a developer and you are using avifile for something check infotypes.h CodecInfo::Get method for more details. As a side effect now the codec order switching works immediately after apply so no restart is needed in this case any more (and if it is - it's a bug and should be reported).
Update of the previous snapshot - this time it's mainly fixing some reported problems - fixes some compilation problems (Mandrake users with gcc-3.1.1 will still have to wait - they use buggy compiler). Also some plugins were not really checked with their new architecture - now they should be at least correctly opened by libaviplay :). Some update have been made in win32 plugin and also streamed asf files shuold not at least display buffer status (the code has been there for a long time but it's been disabled or unused - so finaly users of slower net connection might have some idea why the player stops...). Some more 'throw' has been removed -> less problems with optimalized compilation and some gcc compilers.
New release is here - featuring some fixes for mjpeg playing, various speedups (I'd have call them zero memcopy technology if I'd be a M$ member). As there has been relatively a lot of modification it's possible, that something which has used to work now has some problems - so if you find some problems - just report them.
Some releatively hard to find problems in win32 loader were fixed as well.
I think it would be good if more users would have start to test CVS version - there are some bugfixes - some of them tries to avoid using exceptions in the running code as some compilers seems to produce buggy optimized code. Also there are some bugfixes for win32 loader which might have caused mysterous coredumps. And also there has been very limited the amount of memory copies thus it's faster... Another not yet fully implemented feature is --enable-fast-build - this should dramaticaly reduce building time. For this moment there is no new CVS snapshot - hopefully there will be one next week.
So there will be now a bit longer delay between updates as the avifile seems to now stabilized somewhat and no major bugs were reported so I think it's about the right time to make first few steps for 0.8 release - this will probably mean some major redesign steps so it's possible the CVS might get sometimes into incorrect stages (but I'll always try to minimalize this as much as I could). (Also keep in mind - avifile-0.7 is still fully source compatible with 0.6 release - so if some software requires 0.6 release then 0.7 might and should be used instead)
So what is new for this release - well we have now hyperfast seeking as some other players started to get very close and some even ahead of avifile so it's time to move ahead again. Some ugly bugs were fixed for PCM audio stream length calculation - I'm always amazed how such bugs could survive for such a long time without notice. Support for playing of some broken streams has been improved. Some junk code removement is in progress so other people who might be interested in how the avifile interaly works might read something what actually makes a sence :). A nice side effect is that avifile now behaves better if Linux kernel uses different HZ setting). As usually most of reported problems were fixed - I just remain that avirecompress is for now considered broken.
Ok Friday seems to be the wrong day for software release so lets try another one. This release has slightly improved Qt autodetection so if there were some people with some weird problem in this part - try this release. Also some there are some bugfixes in avi parsing. For this release the behaviour of PgUp/Down CursorUp/Down has been switched and for now it can not be configured.
Looks like it will be the endless stream of bugs all the time so I'm a bit resignating on creation of bugless version :) and I'll now focus back on the support of some new features. Of course bugs will be always eliminated if they will be reported, but the new ones somehow always find its way in. (But usually all reported bugs are fixed within a week).
So for this release there are few new things - support of PIM1 fourcc via ffmpeg. XviD and DivX4 codecs have now slightly improved support for option selection (If some XviD developer reads this - please try to be a bit more conservative with your header file modifications) In this version there are made some experiments with SDL and OSS audio timing - thus if you experience any asynch behaviour - please report so it could be fixed for your box as well. Qt3.0 and font selection should be working and compilation with gcc3.1 as well. Also because avirecompress will support some new stuff it's currently marked as broken.
I guess the amount of bugs has decreased to the level where I could start to think about new features and new directions in developement. It's obvious that there are probably all C++ developers working for KDE so there is no one left for avifile :( thus the future work will be oriented into tight bounding around ffmpeg where the situation is much better.
Highlights of this release - fixed obscure bug with forgotten unrenamed define which has caused some problems with audio renderer selection. Easy fix is to install new avifile and removing your ~/.avirc. Also as plugins no longer contain prefix 'lib' with avifile instalation into the same location user will get twice as much plugins which is not really useful so I suggest to remove previously installed avifile0.7 library subdir before new instalation is made. (Also for less experienced users - there are packages for most common Linux distributions)
Ok - let's try for the third time. Though all reported problems were solved some new ones may always appear. More especially when such major revision of of automake scripts have been made. So it should give much better Qt detection which however might fail for some cases - so if you still have some problem and your instalation is correct please provide some detailed information about installed version and config.log and configure log files.
Here we go with another one - this time it mainly fixes some bugs reported by several user. For users which are using hw acccelareted rendering there is now support for modification of XV attributes so the user might change independently Brigthness Contrast and few other things. There are some known problems i.e. NVidia users could use this feature only with very latest SDL library (or they could recompile the older one with commenting out #define USE_LAST_ADAPTOR in src/video/x11/SDL_x11yuv.c. Some ATI cards seems to expose more attributes then actually supports. Matrox card seems to be just fine, but has the smallest amount of supported attributes (NVidia seems to be the king here).
Other fixes includes support for --without-x (has broken for quite some time). If there are some other non-working option please report. If there is still some Potato user - he might be pleased to know that aviplay know compiles with old Qt2.0 & SDL1.0.0, but such users should be aware of the fact that newer XFree, SDL, Qt are far more powerfull and the video will be in most cases much better.
Ok let's try to make another release - this time it is a test 0.7 version (ok I know there was never final 0.6 - but several users expresed wishes to increase the version number so here we go :)). This release should contain new buffering code, many major and minor bugfixes. Also there is support for DivX5 - DX50 avi compressed video. If you get any problems with this library or player please report your problem (and I just remain that reporting them on non avifile mailing list won't help you).
Notice for users who wants to use the latest unstable automake1.6. This tool has some serious bugs inside - thus if you use it - do not report your problems to avifile list, but rather directly to automake developers. Suggested solution is to downgrade to automake1.5. Also if you are using CVS snahshot you do NOT need to use autogen.sh as configure script is already included in the archive - so just use it directly.
Many stupid bugs were hopefully removed so the player should mostly play everything it used to play. If you still find some file which is unplayable - just report your problem. Hopefully next week there could be released new CVS snapshost.
Let's start to use CVS again. Known problems: Autoquality is still not working. There are some issues with codec setting I know about.
All other problems are bugs that should be reported - especialy unplayble files (also those network streamed - keep in mind there are some network streams which are comming together with some advertisment movie at the begining and the end - so far the player doesn't handle this situation correctly - something like playlist is still missing :(
If user is experiencing interrupts in played network stream - just press press pause and wait for a few secs and press play again (will be handled automaticaly later)
There is slightly updated snaphost from 20/12/2001 in download section, where admin dir has been added. I'd like to put BIG WARNING HERE - please do NOT use current CVS - it's broken and will stay that way for at least another week. There are some serious internal modification being made and it's currently impossible to make them in consistent way so the player will coredump instantly - instead feel free to use the lastest available CVS snaphost.
Most probably the last snapshot in this year. It has couple of new features like support for reading TV stations for .xawtv program and slightly reworked writing classes. I guess until the next year there will be quite limited support from avifile side - so just report your problem and expect solution next year.
CVS version has more robust mp3 detection (no snapshot today). Finaly some bitrate settings should be fully respected by win32 DivX codecs as might be checked in avirecompress. However as there are some unfinished updates in the sources I've considered to not release snaphot as it might introduce know bugs - something I do not want to do - to be more specific segmentation is currently disabled. On the other hand avirecompress now at least writes nice stats during compression - but of course it is very slow as it is not yet using YUV capabilities of many codecs - so if you are looking for some real program avirecompress is probably not the thing you are looking for (but if you want to help with development or even just with managing web pages or writting doc you are always welcome - I guess all the users should read article from abiword developers listed on slashdot recently)
The problem with RH distribution seems to be that some users are probably using some experimental glibc packages - so simply do not use them.
This release has some fixes for dropping and autoquality so it should behave much better on really slow machines (yeah enjoy those last few days when avifile is still the fastest AVI player around here ;)) - and again - if you will not complain I will hardly notice such problems and if you will complain on completely unrelated list to avifile it's even worse..
Also it looks like we now have some problems with RH7.2 and SDL-1.2.3. So if any producer of this system knows the reason why libesd can't find atexit - let us know...