AVC3 content is now rejected

RESOLVED FIXED in Firefox 42

Status

()

defect
RESOLVED FIXED
4 years ago
4 years ago

People

(Reporter: jya, Assigned: jya)

Tracking

(Blocks 1 bug, {regression})

43 Branch
mozilla43
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(firefox42 fixed, firefox43 fixed)

Details

Attachments

(1 attachment)

Assignee

Description

4 years ago
This is due to bug 1202677

It errors if an avcC atom has a size <= 7.

But an AVC3 avcC atom (which has a an empty SPS and PPS) will typically have a size of 7.

This causes all BBC streams to fail
Assignee

Comment 1

4 years ago
An AVC3 stream will typically use 7 bytes avcC (due to have 0 SPS/PPS)
Attachment #8661656 - Flags: review?(gsquelart)
Assignee

Updated

4 years ago
Assignee: nobody → jyavenard
Attachment #8661656 - Flags: review?(gsquelart) → review+
Assignee

Comment 2

4 years ago
Comment on attachment 8661656 [details] [diff] [review]
[mp4] Don't reject 7 bytes long avcC atom.

Approval Request Comment
[Feature/regressing bug #]: 1202677
[User impact if declined]: AVC3 (BBC and others) won't play
[Describe test coverage new/current, TreeHerder]: local
[Risks and why]: None. This is an addendum to an earlier patch that introduced a regression
[String/UUID change made/needed]: None

It's trivial enough that I'm tempted to push it straight to aurora and central!
Attachment #8661656 - Flags: approval-mozilla-aurora?

Comment 3

4 years ago
> It's trivial enough that I'm tempted to push it straight to aurora and central!

It would be awesome if this was backported now we are on the whitelist for 41.
Assignee

Comment 4

4 years ago
the fix that introduced the regression was only uplifted to aurora (42) a few days ago. 41 isn't affected

Comment 5

4 years ago
Great news, thanks.
https://hg.mozilla.org/mozilla-central/rev/95bac0001e70
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla43
Assignee

Comment 8

4 years ago
This has regressed many MSE streams.. Can we get this in in priority?
Blocks: MSE, 1197083
Comment on attachment 8661656 [details] [diff] [review]
[mp4] Don't reject 7 bytes long avcC atom.

Important regression, taking it.
Attachment #8661656 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
You need to log in before you can comment on or make changes to this bug.