Closed
Bug 1245463
Opened 8 years ago
Closed 8 years ago
crash in mp4_demuxer::MP4Metadata::GetNumberTracks, abort should let appendBuffer finish
Categories
(Core :: Audio/Video: Playback, defect, P1)
Tracking
()
RESOLVED
FIXED
mozilla47
People
(Reporter: mozbugz, Assigned: jya)
References
Details
(Keywords: crash)
Crash Data
Attachments
(4 files, 5 obsolete files)
58 bytes,
text/x-review-board-request
|
mozbugz
:
review+
Sylvestre
:
approval-mozilla-aurora+
|
Details |
58 bytes,
text/x-review-board-request
|
mozbugz
:
review+
Sylvestre
:
approval-mozilla-aurora+
Sylvestre
:
approval-mozilla-beta+
|
Details |
58 bytes,
text/x-review-board-request
|
mozbugz
:
review+
Sylvestre
:
approval-mozilla-aurora+
|
Details |
58 bytes,
text/x-review-board-request
|
mozbugz
:
review+
Sylvestre
:
approval-mozilla-aurora+
|
Details |
Follow-up to bug 1239983. The patch there just adds a few diagnostic checks, which should help narrow down the root cause, or disprove some guesses (e.g. bug 1239983 comment 2). NI:myself to regularly monitor new crashes that may happen in FF47 over the next few days. Crash signatures may be a bit different, the following TrackBuffersManager methods should be looked for: ShutdownDemuxers, OnDemuxerResetDone, OnDemuxerInitDone. [Tracking Requested - why for this release]: This seems like a common enough crash. We are investigating.
Flags: needinfo?(gsquelart)
Assignee | ||
Comment 1•8 years ago
|
||
I believe the issue is a race between when the demuxer init promise gets resolved (and queue a task) and when the ResetParserState task is queued. If the ResetParserState task is queued prior the demuxer resolving, then the input buffer will be cleared and the demuxer destroyed ; resulting in the crash reported. I will make sourcebuffer.abort() block until the SegmentParserLoop has completed its run.
Assignee: gsquelart → jyavenard
Flags: needinfo?(gsquelart)
Reporter | ||
Comment 2•8 years ago
|
||
Keeping NI:me as reminder to monitor crash reports, just in case there is useful information there.
Flags: needinfo?(gsquelart)
Assignee | ||
Comment 3•8 years ago
|
||
The changes that follow may cause the active sourcebuffer list to be modified; which will trigger an assertion if the mediasource object is closed. Review commit: https://reviewboard.mozilla.org/r/34311/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34311/
Attachment #8717783 -
Flags: review?(gsquelart)
Assignee | ||
Comment 4•8 years ago
|
||
The W3C spec indicates that while everything in MSE is asynchronous, the abort() command is to interrupt the current segment parser loop and have the reset parser loop synchronously completes the frames present in the input buffer. This causes a fundamental issue that abort() will never result in a deterministic outcome as the segment parser loop may be in different condition. We used to really attempt to abort the current operation, however there could have been a race in the order in which tasks were queued. As such, we now simply wait for the current appendBuffer to complete. This also simplifies the code greatly, as we don't need to worry about pending concurrent appendBuffer. The actually happens to be similar to the Chromium behavior. Review commit: https://reviewboard.mozilla.org/r/34313/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34313/
Attachment #8717784 -
Flags: review?(gsquelart)
Assignee | ||
Comment 5•8 years ago
|
||
This is similar to bug 1239983, we strongly assert should a segment parser loop be running when it must have completed. Review commit: https://reviewboard.mozilla.org/r/34315/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34315/
Attachment #8717785 -
Flags: review?(gsquelart)
Assignee | ||
Comment 6•8 years ago
|
||
Review commit: https://reviewboard.mozilla.org/r/34317/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34317/
Attachment #8717786 -
Flags: review?(gsquelart)
Assignee | ||
Comment 7•8 years ago
|
||
Review commit: https://reviewboard.mozilla.org/r/34319/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34319/
Attachment #8717787 -
Flags: review?(gsquelart)
Reporter | ||
Updated•8 years ago
|
Attachment #8717783 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 8•8 years ago
|
||
Comment on attachment 8717783 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r?gerald https://reviewboard.mozilla.org/r/34311/#review31005
Reporter | ||
Comment 9•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald https://reviewboard.mozilla.org/r/34313/#review31007 r+ with nits: ::: dom/media/mediasource/SourceBuffer.cpp:217 (Diff revision 1) > + // mAborting will become false once the AppendPromise will be resolved. "once the AppendPromise is resolved." ::: dom/media/mediasource/TrackBuffersManager.cpp:160 (Diff revision 1) > MOZ_ASSERT(!mAppendRunning, "AbortAppendData must have been called"); 'AbortAppendData' is dead, so the assert comment is now incorrect. Not sure about the assertion itself, I think it is still needed.
Attachment #8717784 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 10•8 years ago
|
||
Comment on attachment 8717785 [details] MozReview Request: Bug 1245463: [MSE] P3. Diags to ensure the Segment Parser Loop isn't running. r?gerald https://reviewboard.mozilla.org/r/34315/#review31009 ::: dom/media/mediasource/TrackBuffersManager.cpp:160 (Diff revision 1) > - MOZ_ASSERT(!mAppendRunning, "AbortAppendData must have been called"); > + MOZ_RELEASE_ASSERT(!mAppendRunning, "Append is running, abort must have been called"); Ignore my nit in P2. :-)
Attachment #8717785 -
Flags: review?(gsquelart) → review+
Reporter | ||
Updated•8 years ago
|
Attachment #8717786 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 11•8 years ago
|
||
Comment on attachment 8717786 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove dead code. r?gerald https://reviewboard.mozilla.org/r/34317/#review31013
Assignee | ||
Comment 12•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34313/diff/1-2/
Assignee | ||
Updated•8 years ago
|
Attachment #8717783 -
Attachment is obsolete: true
Assignee | ||
Updated•8 years ago
|
Attachment #8717785 -
Attachment is obsolete: true
Assignee | ||
Updated•8 years ago
|
Attachment #8717786 -
Attachment is obsolete: true
Assignee | ||
Updated•8 years ago
|
Attachment #8717787 -
Attachment is obsolete: true
Attachment #8717787 -
Flags: review?(gsquelart)
Assignee | ||
Comment 13•8 years ago
|
||
The changes that follow may cause the active sourcebuffer list to be modified; which will trigger an assertion if the mediasource object is closed. Review commit: https://reviewboard.mozilla.org/r/34325/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34325/
Attachment #8717791 -
Flags: review?(gsquelart)
Assignee | ||
Comment 14•8 years ago
|
||
This is similar to bug 1239983, we strongly assert should a segment parser loop be running when it must have completed. Review commit: https://reviewboard.mozilla.org/r/34327/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34327/
Attachment #8717792 -
Flags: review?(gsquelart)
Assignee | ||
Comment 15•8 years ago
|
||
Review commit: https://reviewboard.mozilla.org/r/34329/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34329/
Attachment #8717793 -
Flags: review?(gsquelart)
Assignee | ||
Comment 16•8 years ago
|
||
Review commit: https://reviewboard.mozilla.org/r/34331/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/34331/
Attachment #8717794 -
Flags: review?(gsquelart)
Reporter | ||
Updated•8 years ago
|
Flags: needinfo?(gsquelart)
Summary: crash in mp4_demuxer::MP4Metadata::GetNumberTracks, follow-up → crash in mp4_demuxer::MP4Metadata::GetNumberTracks, abort should let appendBuffer finish
Reporter | ||
Comment 17•8 years ago
|
||
Comment on attachment 8717791 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald https://reviewboard.mozilla.org/r/34325/#review31017
Attachment #8717791 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 18•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald https://reviewboard.mozilla.org/r/34327/#review31019
Attachment #8717792 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 19•8 years ago
|
||
Comment on attachment 8717793 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove dead code. r?gerald https://reviewboard.mozilla.org/r/34329/#review31021
Attachment #8717793 -
Flags: review?(gsquelart) → review+
Reporter | ||
Comment 20•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald https://reviewboard.mozilla.org/r/34331/#review31023
Attachment #8717794 -
Flags: review?(gsquelart) → review+
Assignee | ||
Comment 21•8 years ago
|
||
Comment on attachment 8717791 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34325/diff/1-2/
Assignee | ||
Comment 22•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34313/diff/2-3/
Assignee | ||
Comment 23•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34327/diff/1-2/
Assignee | ||
Comment 24•8 years ago
|
||
Comment on attachment 8717793 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove dead code. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34329/diff/1-2/
Assignee | ||
Comment 25•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34331/diff/1-2/
Assignee | ||
Updated•8 years ago
|
status-firefox44:
--- → wontfix
status-firefox45:
--- → affected
status-firefox46:
--- → affected
status-firefox47:
--- → affected
Assignee | ||
Comment 26•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34313/diff/3-4/
Assignee | ||
Comment 27•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34327/diff/2-3/
Assignee | ||
Comment 28•8 years ago
|
||
Comment on attachment 8717793 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove dead code. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34329/diff/2-3/
Assignee | ||
Comment 29•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34331/diff/2-3/
Comment 30•8 years ago
|
||
https://hg.mozilla.org/integration/mozilla-inbound/rev/5b37030dbfe3 https://hg.mozilla.org/integration/mozilla-inbound/rev/a9ccf315f1f2 https://hg.mozilla.org/integration/mozilla-inbound/rev/2376ec200aec https://hg.mozilla.org/integration/mozilla-inbound/rev/2affd4506079 https://hg.mozilla.org/integration/mozilla-inbound/rev/22641351f11f
Comment 31•8 years ago
|
||
Backout: https://hg.mozilla.org/integration/mozilla-inbound/rev/a2c24c4351b4 https://hg.mozilla.org/integration/mozilla-inbound/rev/0e2caa2fb97e https://hg.mozilla.org/integration/mozilla-inbound/rev/1939e6b7e789 https://hg.mozilla.org/integration/mozilla-inbound/rev/eefd40bb4219 https://hg.mozilla.org/integration/mozilla-inbound/rev/c2253ed9eaef
Comment 32•8 years ago
|
||
sorry backout for test failures like https://treeherder.mozilla.org/logviewer.html#?job_id=21417727&repo=mozilla-inbound
Flags: needinfo?(jyavenard)
Comment 33•8 years ago
|
||
btw seems https://treeherder.mozilla.org/logviewer.html#?job_id=21422960&repo=mozilla-inbound is also related
Assignee | ||
Updated•8 years ago
|
Flags: needinfo?(jyavenard)
Assignee | ||
Comment 34•8 years ago
|
||
Comment on attachment 8717791 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34325/diff/2-3/
Attachment #8717791 -
Attachment description: MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r?gerald → MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald
Assignee | ||
Comment 35•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34313/diff/4-5/
Attachment #8717784 -
Attachment description: MozReview Request: Bug 1245463: [MSE] P2. When abort() is called, wait until the current appendBuffer completes. r?gerald → MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald
Assignee | ||
Comment 36•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34327/diff/3-4/
Attachment #8717792 -
Attachment description: MozReview Request: Bug 1245463: [MSE] P3. Diags to ensure the Segment Parser Loop isn't running. r?gerald → MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r=gerald
Assignee | ||
Comment 37•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34331/diff/3-4/
Attachment #8717794 -
Attachment description: MozReview Request: Bug 1245463: [MSE] P5. Remove no longer working Dump() commands. r?gerald → MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald
Assignee | ||
Updated•8 years ago
|
Attachment #8717793 -
Attachment is obsolete: true
Reporter | ||
Comment 38•8 years ago
|
||
https://reviewboard.mozilla.org/r/34313/#review31281
Reporter | ||
Comment 39•8 years ago
|
||
https://reviewboard.mozilla.org/r/34327/#review31283 ::: dom/media/mediasource/TrackBuffersManager.cpp:611 (Diff revision 4) > + mAppendRunning = true; Is this one needed?
Reporter | ||
Comment 40•8 years ago
|
||
https://reviewboard.mozilla.org/r/34331/#review31285
Assignee | ||
Comment 41•8 years ago
|
||
(In reply to Gerald Squelart [:gerald] from comment #39) > https://reviewboard.mozilla.org/r/34327/#review31283 > > ::: dom/media/mediasource/TrackBuffersManager.cpp:611 > (Diff revision 4) > > + mAppendRunning = true; > > Is this one needed? no... bad rebase.
Assignee | ||
Comment 42•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34327/diff/4-5/
Assignee | ||
Comment 43•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34331/diff/4-5/
Assignee | ||
Comment 44•8 years ago
|
||
Comment on attachment 8717784 [details] MozReview Request: Bug 1245463: [MSE] P2. Remove MediaSource's duration mirror. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34313/diff/5-6/
Assignee | ||
Comment 45•8 years ago
|
||
Comment on attachment 8717792 [details] MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34327/diff/5-6/
Attachment #8717792 -
Attachment description: MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r=gerald → MozReview Request: Bug 1245463: [MSE] P3. When abort() is called, wait until the current appendBuffer completes. r?gerald
Assignee | ||
Comment 46•8 years ago
|
||
Comment on attachment 8717794 [details] MozReview Request: Bug 1245463: [MSE] P4. Remove no longer working Dump() commands. r=gerald Review request updated; see interdiff: https://reviewboard.mozilla.org/r/34331/diff/5-6/
Comment 47•8 years ago
|
||
https://hg.mozilla.org/integration/mozilla-inbound/rev/48364b868055 https://hg.mozilla.org/integration/mozilla-inbound/rev/e8bb35ea72e4 https://hg.mozilla.org/integration/mozilla-inbound/rev/83a2de2fcf02 https://hg.mozilla.org/integration/mozilla-inbound/rev/17ffbe7ca24e
Comment 48•8 years ago
|
||
https://hg.mozilla.org/integration/mozilla-inbound/rev/dca0577d741c https://hg.mozilla.org/integration/mozilla-inbound/rev/75d96e417aaf
Comment 49•8 years ago
|
||
Backed out in https://hg.mozilla.org/integration/mozilla-inbound/rev/8cdb558a60bb for assertions like https://treeherder.mozilla.org/logviewer.html#?job_id=21611746&repo=mozilla-inbound persisting past the followups.
Comment 51•8 years ago
|
||
https://hg.mozilla.org/integration/mozilla-inbound/rev/4c9a6c4c5ba7 https://hg.mozilla.org/integration/mozilla-inbound/rev/39d6a90be824 https://hg.mozilla.org/integration/mozilla-inbound/rev/f97951c80d17 https://hg.mozilla.org/integration/mozilla-inbound/rev/c0c83da3ec2e https://hg.mozilla.org/integration/mozilla-inbound/rev/48cd4633a510
Comment 52•8 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/4c9a6c4c5ba7 https://hg.mozilla.org/mozilla-central/rev/39d6a90be824 https://hg.mozilla.org/mozilla-central/rev/f97951c80d17 https://hg.mozilla.org/mozilla-central/rev/c0c83da3ec2e https://hg.mozilla.org/mozilla-central/rev/48cd4633a510
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla47
Assignee | ||
Comment 53•8 years ago
|
||
Comment on attachment 8717791 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald Approval is for all patches. Approval Request Comment [Feature/regressing bug #]: 1245463 [User impact if declined]: Crashes, YouTube stalls [Describe test coverage new/current, TreeHerder]: Manual test, lots of mochitest and regression tests (and hence why this was backed out a few times until it was perfect) [Risks and why]: Medium. We introduce a wait on the main thread which has historically been a problem in the past resulting in deadlocks (though the code deadlocking was always wrong). However, we are implementing the specs to the letter (which also requires a synchronous wait). Great care was taken to ensure that this wouldn't happen. We have added lots of release assert that will immediately be triggered should this situation happen, allowing for noticing very quickly and should come to worse: back it out or fix it quickly. [String/UUID change made/needed]: None
Attachment #8717791 -
Flags: approval-mozilla-beta?
Attachment #8717791 -
Flags: approval-mozilla-aurora?
Comment 54•8 years ago
|
||
Comment on attachment 8717791 [details] MozReview Request: Bug 1245463: [MSE] P1. Do not attempt to retrieve the buffered range if the mediasource is in closed state. r=gerald This crash already impacts the release and the number of crashes on all channels is low. As the patches are pretty big and the risk medium, I prefer that this rides the train from aurora.
Attachment #8717791 -
Flags: approval-mozilla-beta?
Attachment #8717791 -
Flags: approval-mozilla-beta+
Attachment #8717791 -
Flags: approval-mozilla-aurora?
Attachment #8717791 -
Flags: approval-mozilla-aurora+
Updated•8 years ago
|
Updated•8 years ago
|
Attachment #8717784 -
Flags: approval-mozilla-aurora+
Updated•8 years ago
|
Attachment #8717792 -
Flags: approval-mozilla-aurora+
Updated•8 years ago
|
Attachment #8717794 -
Flags: approval-mozilla-aurora+
Assignee | ||
Comment 55•8 years ago
|
||
I was about to ask that only P1 be beta+ then, but I see that you've approved it for this one which is great (bug 1247875). This one is for a new feature only in 45 and later, and my guess is that once 45 becomes release, it's severity will increase greatly.
Comment 56•8 years ago
|
||
Sorry, I wasn't planning to accept in beta but it seems that my mistake was in fact correct :p
Reporter | ||
Comment 57•8 years ago
|
||
Good news, everyone! A crash has been induced by the patch in bug 1239983, as hoped: https://crash-stats.mozilla.com/report/index/b16687f2-026c-439d-a408-aa4fc2160213 It shows that an unexpected CompleteResetParserState() happens while an init/reset Promise is in flight. Hopefully Jean-Yves' patch should rectify this situation. I'll keep monitoring crashes, if all goes well we might see a few from builds that contain bug 1239983 (> 2016-02-03), and then no more for builds containing patches from this bug (> 2016-02-14).
Crash Signature: [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ] → [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ]
[@ mozilla::TrackBuffersManager::ShutdownDemuxers ]
Comment 58•8 years ago
|
||
bugherder uplift |
https://hg.mozilla.org/releases/mozilla-aurora/rev/8548a5072e90 https://hg.mozilla.org/releases/mozilla-aurora/rev/4ebe7c8ee824 https://hg.mozilla.org/releases/mozilla-aurora/rev/6a94041dcf4e https://hg.mozilla.org/releases/mozilla-aurora/rev/67a52be76498 https://hg.mozilla.org/releases/mozilla-aurora/rev/9220005ce4c9
Comment 59•8 years ago
|
||
https://hg.mozilla.org/releases/mozilla-beta/rev/123cfb0366b7 https://hg.mozilla.org/releases/mozilla-beta/rev/09b2d06b4ed0 https://hg.mozilla.org/releases/mozilla-beta/rev/317fdb031d06 https://hg.mozilla.org/releases/mozilla-beta/rev/d243af3941f8 https://hg.mozilla.org/releases/mozilla-beta/rev/929c6a005171
Assignee | ||
Comment 60•8 years ago
|
||
Sylvestre, looks like uplift has been too eager and the lot has been uplift d to beta.. Rather than revert the lot, what about we let it bake for a week and see how it goes?
Flags: needinfo?(sledru)
Comment 61•8 years ago
|
||
Oups :( Well, we built beta 5 with it. Looks like we don't have much choice anyway :p
Flags: needinfo?(sledru)
Updated•8 years ago
|
Crash Signature: [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ]
[@ mozilla::TrackBuffersManager::ShutdownDemuxers ] → [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ]
[@ mozilla::TrackBuffersManager::ShutdownDemuxers ]
[@ mozilla::dom::MediaSource::Duration]
Comment 63•8 years ago
|
||
this is marked fixed in 45, but the mozilla::dom::MediaSource::Duration signature is one of the top signatures in 45.0.1 & making up 3% of all crashes there atm... should we reopen this bug/file a new issue?
Flags: needinfo?(jyavenard)
Assignee | ||
Comment 64•8 years ago
|
||
I don't believe the two are related (mozilla::dom::MediaSource::Duration) has nothing to do with the null deref in mp4_demuxer::
Flags: needinfo?(jyavenard)
Assignee | ||
Updated•8 years ago
|
Crash Signature: [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ]
[@ mozilla::TrackBuffersManager::ShutdownDemuxers ]
[@ mozilla::dom::MediaSource::Duration] → [@ mp4_demuxer::MP4Metadata::GetNumberTracks ]
[@ mp4_demuxer::MP4Metadata::GetNumberTracks const ]
[@ mozilla::TrackBuffersManager::ShutdownDemuxers ]
You need to log in
before you can comment on or make changes to this bug.
Description
•