dcpomatic/test, branch v2.17.12 DCP-o-matic DCP tools https://git.carlh.net/cgit/dcpomatic/atom?h=v2.17.12 2024-02-22T00:25:03Z Merge tag 'v2.16.78' into v2.17.x 2024-02-22T00:25:03Z Carl Hetherington cth@carlh.net 2024-02-22T00:25:03Z urn:sha1:a45f6245e8cc785fab436c9282fa3d1baf3a8575 Make DCPExaminer::size() optional and deal with the consequences. 2024-02-21T17:48:53Z Carl Hetherington cth@carlh.net 2024-02-21T09:47:38Z urn:sha1:3ffd0163026be24e5373e0674c3301ed37546e44 This means we can fix the case of a VF having no known size in a nice way, in turn fixing problems caused by the fix to #2775. Fix failure to examine non-flat-ratio VFs (#2775). 2024-02-21T00:27:31Z Carl Hetherington cth@carlh.net 2024-02-21T00:10:23Z urn:sha1:a9b1c1cb65e1902a64430977cf698054e131a6f4 Merge tag 'v2.16.77' into v2.17.x 2024-02-19T22:22:31Z Carl Hetherington cth@carlh.net 2024-02-19T22:22:31Z urn:sha1:51dda295b5bd63e6635ae17786e4657c41a3678a Reset _next_video_time when deciding to emit black. 2024-02-19T08:54:43Z Carl Hetherington cth@carlh.net 2024-02-18T22:36:10Z urn:sha1:167ff6413dcc32680b2b778303f659656c2a51df Otherwise if we inaccurately seek into an area with no video there will never be any content coming into ::video() which could reset _next_video_time. Then when a black frame is emitted using emit_video_until we will emit black frames from time 0 to the time of the inaccurate seek. Add Unzipper. 2024-02-16T22:37:59Z Carl Hetherington cth@carlh.net 2024-02-15T00:05:12Z urn:sha1:befbada161a66a16f8ced89773ad11c0ac5c91e5 Merge branch 'main' into v2.17.x 2024-02-11T23:06:13Z Carl Hetherington cth@carlh.net 2024-02-11T23:06:13Z urn:sha1:6f9adc9f7215a362d23e45b861017638ec67c723 Don't add a font with an empty ID for ccaps either (#2762). 2024-02-11T22:31:39Z Carl Hetherington cth@carlh.net 2024-02-11T19:13:04Z urn:sha1:eb04ac87ccfa046dd342ca7b9e6478c3bdcabbba Previously this was all changed so that for subs we add the default font in a cleaner way (in db22f81ccce9e1a5f205e6d8b3c0631fc039a173). I can't see why we shouldn't treat ccaps the same way, and indeed not doing it causes an error as shown by the included test. Fix apparently ineffective processors when analysing audio (#2671). 2024-02-11T21:55:27Z Carl Hetherington cth@carlh.net 2024-02-11T12:10:48Z urn:sha1:527093fa643049d370f4daee8206349981edbf95 A whole film ("DCP side") analysis would behave the same as a single piece of content analysis if the film only had one piece of content. Here we also clarify that audio analysis of the whole film is different to that for a single piece of content; for the whole film we use processors, and for single pieces of content we do not. Remove apparently out-of-date comment. 2024-02-06T23:21:44Z Carl Hetherington cth@carlh.net 2024-02-06T10:07:53Z urn:sha1:fbcfa342d9f8d6f1430065ad18a1a1b8f7b925db