Changelog for moc-2.6_alpha3+r3005-lp151.7.3.x86_64.rpm :

* Tue Sep 01 2020 obs-service-tar-scmAATTinvalid- Update to version 2.6_alpha3+r3005:
* Account for zero-sample channels produced by some AAC encoding.
*
* When encoding stereo audios, some channel mappings used produce
* multichannel files with the samples of the additional channels
* being all zero. This upsets MOC\'s duration estimation so zero
* valued samples are now ignored.
* Avoid unnecessary downmixing when only determining audio duration.
*
* For the purposes of estimating the duration, we don\'t need to incur
* the overhead of downmixing 5.1 audios.
* Fix: Library downmixed AAC audios produce whitenoise when played.
*
* MOC asks the FAAD2 library to downmix 5.1 audios to stereo, but MOC
* did not adjust the channel count from that of the original audio.
* Fix: Unshifted keypad ENTER is not treated as keyboard ENTER.
*
* This problem arises when a (typically GUI) user expects that the
* ENTER key on the numeric keypad should work the same as the ENTER key
* on the keyboard. This is a reasonable expectation, but it doesn\'t.
*
* How the keypad ENTER key is reported to the application depends on
* the termcap \'AATT8\' or terminfo \'kent\' capability. Using \'linux\' as the
* $TERM setting (as consoles usually do) results in the default 0x0a
* (newline) being reported for both ENTER keypresses, while \'xterm\' (and
* other GUI settings) mostly return newline for keyboard ENTER but 0x157
* (represented by ncurses as KEY_ENTER) for keypad ENTER.
*
* (The shifted keypad ENTER does return a newline in all cases.)
*
* Of the $TERM settings tested, this commit fixes the problem for all
* but \'xterm-color\' which returns 0xcf. However, \'xterm-color\' may
* have been superceded by \'xterm-16color\' and \'xterm-256color\' anyway.
* Fix: Misaligned MOCP_POPTRC manpage paragraph.
*
* Amends: r2734
* Fix: SIGABRT while seeking when file has negative start time.
*
* Subtracting a negative start time from INT64_MAX results in an
* integer overflow.
*
* A negative start time can arise when a file has been cut from a
* longer video file and reference frames prior to the cut point need
* to be included in order to correctly decode the subsequent frames.
*
* Thanks to: Joan Bruguera Mico <joanbrugueramAATTgmail.com>
* Resolves: node/2296 (Seek doesn\'t work with negative start time)
* Fix: Delayed reporting of \"TERMINAL TOO SMALL\".
*
* It seems that performing a temporary suspend (via endwin() /
* refresh()) is the accepted way to force Curses to refresh its
* terminal context, and it isn\'t until after that that the new COLS
* and LINES settings take effect. It feels uncomfortable to have
* that update happen as a side-effect of an otherwise redundant
* suspend sequence, however, it appears to be the only portable way
* of doing it.
*
* Multiple refresh()s may introduce some visible screen flicker (which
* the code elsewhere in MOC tries to avoid), but it has not been
* observed in testing and is probably masked by the resizing operation.
*
* Thanks to: Joan Bruguera Mico <joanbrugueramAATTgmail.com>
* Resolves: node/2177 (MOC prints \"TERMINAL TOO SMALL\" when it\'s not)
* Clean: Rename header guard to avoid potential future conflicts.
*
* Thanks to: Vadim Zhukov <zhukAATTopenbsd.org>
* Clean: Upgrade warning option to GCC\'s more descriptive replacement.
* Clean: Use xcalloc() to obtain cleared memory.
* Tue Jul 04 2017 and.novemberAATTopensuse.org- use NcursesW (wide character support)