dcpomatic/src, branch v2.15.32 DCP-o-matic DCP tools https://git.carlh.net/cgit/dcpomatic/atom?h=v2.15.32 2019-11-11T13:15:44Z Don't trust video timestamps from FFmpegDecoder. 2019-11-11T13:15:44Z Carl Hetherington cth@carlh.net 2019-11-10T21:59:39Z urn:sha1:98342fb53eae4d32440fc69c279f2ca0fef785b5 Make DecoderPart::_position an optional. 2019-11-11T13:15:44Z Carl Hetherington cth@carlh.net 2019-11-10T21:51:55Z urn:sha1:0e8a1ab7c41756115f44229053e1e7024530fb32 Fix comment. 2019-11-11T13:15:19Z Carl Hetherington cth@carlh.net 2019-11-11T13:15:05Z urn:sha1:c247a1b81b78164acf84f4b003062fab25e64b53 Try to fix no-stretch / no-scale for non-square pixels (#1636). 2019-11-08T16:21:19Z Carl Hetherington cth@carlh.net 2019-11-08T16:21:19Z urn:sha1:ee0f2219f3799881bc9f5060edd2ceeecff4217d Remove duplicate log introduced by previos. 2019-11-06T21:28:44Z Carl Hetherington cth@carlh.net 2019-11-06T21:28:44Z urn:sha1:bb285dedfb798ddaada80fdfdddf5873cb990289 More logging when finding assets to put in reels. 2019-11-06T21:28:09Z Carl Hetherington cth@carlh.net 2019-11-06T21:27:04Z urn:sha1:01eb94548560229dc9f633fff7d0602f04ae052c Try to ignore exceptions thrown from boost::filesystem::is_symlink (#1662). 2019-11-06T12:48:37Z Carl Hetherington cth@carlh.net 2019-11-06T12:48:17Z urn:sha1:c55763e650b12b96b97ef8630d3a5277c1237069 Differentiate between stop and suspend in player. 2019-11-05T22:02:59Z Carl Hetherington cth@carlh.net 2019-11-05T22:01:24Z urn:sha1:c8db1d4012217ec295d2a77581cf0661543dc9af This and the previous commit somewhat improve the previously terrible behaviour when seeking during playback, probably introduced by the idle update stuff. Don't request idle get if we are starting playback again. 2019-11-05T20:20:32Z Carl Hetherington cth@carlh.net 2019-11-05T20:20:32Z urn:sha1:25db72d238179045bf5dcc3dbe87658dd3609d52 Remove dubious _buffers_mutex and maintain a lock on _mutex for 2019-11-05T20:09:37Z Carl Hetherington cth@carlh.net 2019-11-05T20:09:37Z urn:sha1:e97d48b043fe39ec22687555225d6b4b526a2172 the whole of ::audio. Otherwise changes to pending seeks can be mixed up with audio being put into the ringbuffer.