The current MediaTime::MaximumTimeScale is 0x7fffffff (2^31-1), which is quite undesirable: it's a prime number, so it does not accomodate *any* other timescale without rounding. I propose to use 1e9 (nanosecond timescale) as maximum timescale: it's very close to 2^31 (a bit under 2^30 actually) and it's already being used by both Matroska (WebM) and GStreamer.
LGTM. The initial choice of maximum time scale was relatively arbitrary; it's less than 1/2 of the maximum possible value of a uint32_t.
Created attachment 352499 [details] Patch
Comment on attachment 352499 [details] Patch Clearing flags on attachment: 352499 Committed r237208: <https://trac.webkit.org/changeset/237208>
All reviewed patches have been landed. Closing bug.
<rdar://problem/45322286>
The change made in https://trac.webkit.org/changeset/237208/webkit to TestWTF.WTF.MediaTime has caused it to fail constantly on all systems. Log: https://build.webkit.org/builders/Apple%20High%20Sierra%20Debug%20WK1%20(Tests)/builds/6021/steps/run-api-tests/logs/stdio Build: https://build.webkit.org/builders/Apple%20High%20Sierra%20Debug%20WK1%20(Tests)/builds/6021
Linked wrong build. Correct build: https://build.webkit.org/builders/Apple%20High%20Sierra%20Release%20WK1%20%28Tests%29/builds/8636
Looks like these tests need rebaselining.
Can this be fixed today?
Re-opened since this is blocked by bug 190691
Yes, the tests need a rebaseline indeed.
Created attachment 352683 [details] Patch
Comment on attachment 352683 [details] Patch Clearing flags on attachment: 352683 Committed r237278: <https://trac.webkit.org/changeset/237278>