Follow-up from https://bugs.webkit.org/show_bug.cgi?id=225630#c9 bug 225630 changed the logic on how we are to reject an append only after one has already been added. One downside of the change was also that it broke the class separation between SourceBuffer and SourceBufferPrivate ; it is preferable to have all the handling of buffer size in the SourceBufferPrivate which avoid ambiguities as discussed in bug https://bugs.webkit.org/show_bug.cgi?id=226034
<rdar://problem/78937909>
Created attachment 430730 [details] Patch
Comment on attachment 430730 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=430730&action=review > Source/WebCore/platform/graphics/SourceBufferPrivate.cpp:726 > + if (WTF::sumOverflows<uint64_t>(totalTrackBufferSizeInBytes(), requiredSize)) > + return true; > + > + return totalTrackBufferSizeInBytes() + requiredSize >= maximumBufferSize; Irrespective of the overall patch, this performs the addition twice. It would probably be more efficient to do something like: uint64_t totalRequired = 0; if (CheckedState::DidOverflow == CheckedSum(totalTrackBufferSizeInBytes(), requiredSize).safeGet(totalRequired); return true; return totalRequired >= maximumBufferSize;
Created attachment 430806 [details] Patch
Comment on attachment 430806 [details] Patch test failures unrelated to media's MSE
jean-yves.avenard@apple.com does not have committer permissions according to https://raw.githubusercontent.com/WebKit/WebKit/main/Tools/Scripts/webkitpy/common/config/contributors.json. Rejecting attachment 430806 [details] from commit queue.
Committed r278603 (238592@main): <https://commits.webkit.org/238592@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 430806 [details].