summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2004-08-03The ultimate solution (like v!@gra? :-)) for 'make dist' and srcdir != builddir.František Dvořák
CVS patchset: 6873 CVS date: 2004/08/03 19:35:11
2004-08-02Forgotten change due to external ffmpeg.František Dvořák
CVS patchset: 6872 CVS date: 2004/08/02 18:41:30
2004-08-02no objections so far: committing the fine speed control patchMiguel Freitas
(in case of major trouble just revert the patch i sent to xine-devel) CVS patchset: 6871 CVS date: 2004/08/02 12:51:07
2004-08-02disable unused functionMichael Roitzsch
CVS patchset: 6870 CVS date: 2004/08/02 12:34:05
2004-08-02configure needs to be run so that the Makefile actually existsMichael Roitzsch
CVS patchset: 6869 CVS date: 2004/08/02 12:32:30
2004-08-02Don't include mmx.h on non-MMX systems: fixes compilation of Goom on Mac OS XAndre Pang
CVS patchset: 6868 CVS date: 2004/08/02 08:25:11
2004-08-02Compile fix when using internal ffmpeg libraryAndre Pang
CVS patchset: 6867 CVS date: 2004/08/02 07:53:52
2004-08-01patch by Robin Kay to dynamically increase the memory stack when it runs fullMichael Roitzsch
CVS patchset: 6866 CVS date: 2004/08/01 16:18:43
2004-07-31New configure option --with-external-ffmpeg[=PREFIX].František Dvořák
CVS patchset: 6865 CVS date: 2004/07/31 18:57:45
2004-07-30possibly fix the sync problem recently reported on xine-develMiguel Freitas
CVS patchset: 6864 CVS date: 2004/07/30 19:08:47
2004-07-29Better fix for 'make dist' with srcdir != builddir.František Dvořák
CVS patchset: 6863 CVS date: 2004/07/29 19:24:50
2004-07-29several fixes to implementation of critical sections using pthreadsMiguel Freitas
(critical section is defined as recursive by MS). strict posix systems will be happier now. workaround a crash with QT dll introduced with the previous commit. CVS patchset: 6862 CVS date: 2004/07/29 18:47:36
2004-07-29mention the win32 dll fixMiguel Freitas
CVS patchset: 6861 CVS date: 2004/07/29 14:48:04
2004-07-29fix from Casper Boemann to reuse the stubs for unresolved symbolsMiguel Freitas
CVS patchset: 6860 CVS date: 2004/07/29 14:45:03
2004-07-27from the docs of sprintf: "when the strings overlap, the behaviour is undefined"Michael Roitzsch
(actually it works with glibc and sprintf, but not with snprintf) CVS patchset: 6859 CVS date: 2004/07/27 18:06:45
2004-07-27* use a header for the init functionsMichael Roitzsch
* do not use void* in pointer arithmetics, since sizeof(void) is undefined -> I converted the pointers to int16_t* since that is common for audio samples in xine and casted them back to uint8_t* for the arithmetics CVS patchset: 6858 CVS date: 2004/07/27 17:59:58
2004-07-26we have time stretching nowMiguel Freitas
CVS patchset: 6857 CVS date: 2004/07/26 22:37:37
2004-07-26play 20% faster by defaultMiguel Freitas
CVS patchset: 6856 CVS date: 2004/07/26 22:34:00
2004-07-26experimental time stretching pluginMiguel Freitas
CVS patchset: 6855 CVS date: 2004/07/26 22:23:34
2004-07-25protect bufferMichael Roitzsch
CVS patchset: 6854 CVS date: 2004/07/25 17:47:01
2004-07-25strncpy will not necessarily terminate the stringMichael Roitzsch
CVS patchset: 6853 CVS date: 2004/07/25 17:45:22
2004-07-25protect buffer; a long configured codec path could have overflown itMichael Roitzsch
CVS patchset: 6852 CVS date: 2004/07/25 17:44:10
2004-07-25the worst case we will write to this buffer is largerMichael Roitzsch
CVS patchset: 6851 CVS date: 2004/07/25 17:42:55
2004-07-25protect some buffers that could overflowMichael Roitzsch
CVS patchset: 6850 CVS date: 2004/07/25 17:42:22
2004-07-25* increase buffer size, "Cseq: %u" has a worst case of 17Michael Roitzsch
* limit TCP port range CVS patchset: 6849 CVS date: 2004/07/25 17:13:54
2004-07-25protect some buffers to be on the safe side (some of them might be hardMichael Roitzsch
to overflow, but it could be possible) input_cdda contains the copy&paste bug of filling the buffer to its end and then appending yet another character CVS patchset: 6848 CVS date: 2004/07/25 17:11:59
2004-07-25a buffer of size 1024 can only be filled with 1023 charactersMichael Roitzsch
CVS patchset: 6847 CVS date: 2004/07/25 17:08:07
2004-07-25protect buffersMichael Roitzsch
CVS patchset: 6846 CVS date: 2004/07/25 17:06:29
2004-07-25do not fill up to the max when you intend to strcat() another character laterMichael Roitzsch
CVS patchset: 6845 CVS date: 2004/07/25 17:06:03
2004-07-25remove some finished tasks and add some new ones for post 1.0Michael Roitzsch
CVS patchset: 6844 CVS date: 2004/07/25 17:04:44
2004-07-24Don't use MAP_FIXED to place registers after video RAM.Robin KAY
CVS patchset: 6843 CVS date: 2004/07/24 17:26:02
2004-07-24When including AC_TRY_COMPILE in the ACTION-IF-NOT-FOUND section of another ↵Stephen Torri
AC_TRY_COMPILE ensure that second call is wrapped in brackets ([ ACE_TRY_COMPILE ...]). Apparently without them autoheader and automake get confused. Replaced the AC_LANG_CPLUSPLUS, AC_LANG_SAVE and AC_LANG_RESTORE with the recommended calls of AC_LANG_PUSH(C++) and AC_LANG_POP([C++]). Let me know if the last two things cause problems. I can revert this change. CVS patchset: 6842 CVS date: 2004/07/24 02:35:47
2004-07-23fix configfile def_value assertion.Daniel Caujolle-Bert
CVS patchset: 6841 CVS date: 2004/07/23 13:47:39
2004-07-23made io_helper.c compile under Solaris 2.6 (socklen_t undefine type) (close ↵Daniel Caujolle-Bert
#992824) CVS patchset: 6840 CVS date: 2004/07/23 12:42:14
2004-07-22protect from overly long filenamesMichael Roitzsch
CVS patchset: 6839 CVS date: 2004/07/22 14:29:28
2004-07-222 extra bytes is enough, one for the '/' and one for '\0'Michael Roitzsch
CVS patchset: 6838 CVS date: 2004/07/22 14:27:48
2004-07-22frontends could have crashed xine-lib by passing a too long filenameMichael Roitzsch
CVS patchset: 6837 CVS date: 2004/07/22 14:26:24
2004-07-22since we potentially add a '\n', we cannot fill the string to the limit earlierMichael Roitzsch
CVS patchset: 6836 CVS date: 2004/07/22 14:25:05
2004-07-22long paths to the vidix drivers would have overrun the path bufferMichael Roitzsch
CVS patchset: 6835 CVS date: 2004/07/22 14:23:42
2004-07-22* font name stored in an unprotected buffer, configuring a long font nameMichael Roitzsch
would have caused an overflow * using strncpy is good, but if the buffer is too short, it leaves the string unterminated; fixed CVS patchset: 6834 CVS date: 2004/07/22 14:21:31
2004-07-22* do not use strcpy with overlapping strings, the behaviour is undefinedMichael Roitzsch
* fixed lots of unprotected buffers, overlong subtitle lines would have caused overflows CVS patchset: 6833 CVS date: 2004/07/22 14:19:12
2004-07-22a broken CD might contain an unterminated label, which would have overrun theMichael Roitzsch
buffer; a label with '%' characters would have lead to a format string attack Rocky, can you merge this back to libcdio, please? CVS patchset: 6832 CVS date: 2004/07/22 14:15:19
2004-07-22do not use functions in array initialization, this is actually not allowed in CMichael Roitzsch
and some compilers do not grok it CVS patchset: 6831 CVS date: 2004/07/22 14:12:20
2004-07-22embarassing overflow in the hacker's guide example appMichael Roitzsch
CVS patchset: 6830 CVS date: 2004/07/22 14:02:49
2004-07-22remove some things which are doneMichael Roitzsch
CVS patchset: 6829 CVS date: 2004/07/22 14:00:58
2004-07-21new goom2k4-dev18 includes almost all of my warning fixesMichael Roitzsch
(the goom team is quite fast in applying my patches :) ) some files are still slightly patched from the release version, but I will get this sorted out so we will soon use a completely unmodified version of goom it compiles almost warningless now (the remaining warnings are introduced by lex/yacc and I plan to simply ignore them) the Makefile.am distclean-hack is also obsolete now, since the offending files have been renamed CVS patchset: 6828 CVS date: 2004/07/21 14:38:29
2004-07-212 bugs fixed actuallyMiguel Freitas
CVS patchset: 6827 CVS date: 2004/07/21 00:59:50
2004-07-20Use lavc version of the mpeg rff flag.Thibaut Mattern
CVS patchset: 6826 CVS date: 2004/07/20 21:38:34
2004-07-20fix compiler warningsMichael Roitzsch
CVS patchset: 6825 CVS date: 2004/07/20 16:39:37
2004-07-20small DXR3 code cleanup:Michael Roitzsch
* make some helper functions non-inline that Bastien had to move because of their inlining (I realized that some helpers are a bit too large for inlining, so they might worsen the I-cache usage.) * properly use logging macros in all DXR3 code * prepare SPU decoder for new (and hopefully correct) SPU forcing; this should fix some longstanding DXR3 SPU problems, but the code needs more testing so it is disabled for now CVS patchset: 6824 CVS date: 2004/07/20 16:37:44