https://github.com/cisco/openh264/commit/032f97db96827f9c4c183de5bc3794fce7b76bf6

lol, I must have pushed submit a seconds before you did. See and feel free to close as dupe my bug 1619986. Would be good to copy over the info I posted there.

For interested parties.

v2.1.0

Experimentally support for multi-thread decoding(default disabled,and may result in random problems if enabled)
Assembly optimization for loongson platform
Update meson version to 5
Some minor bug fixes

Will this finally make it into FF78?

Looks like 2.1.1 is now released since May.

This would be nice for FF80.

For interested parties

v2.1.1

release x86 and x64 libraries for android platform
Bug fixes

So what's the holdup for getting this into FF? Surely this is worth getting into 82?

A note that this would allow to use OpenH264 GMP as fallback for H264 for playing normal video files, as OpenH264 now supports High profile

Unfortunately, I don't have time to work on this at the moment.

Please be sure to check oss-fuzz[1] before updating.

[1] https://bugs.chromium.org/p/oss-fuzz/issues/list?q=openh264

Still not worth adding yet?

gmpopenh264.dll which is installed by Firefox is 1.8.1.1 and the binary is signed by Mozilla in 2019-03-03.
Are there any security issues fixed between 1.8.1.1 and 2.2.0?

(In reply to Takanori MATSUURA from comment #13)

Are there any security issues fixed between 1.8.1.1 and 2.2.0?

The issue from Comment 10 appears to be resolved. Really, there's no good excuse not to bring this component to the current v2.2.0.

But it might also be good to see if https://github.com/cisco/openh264/pull/3477 gets merged and they push out a v2.2.1 release as PR3477 seems important.

Still no movement on PR3477 but at least PR3478 is fixed. Let's hope it's not another nearly 2 years before they release 2.2.1.

Dear Triage Owner,

OpenH264 plugin which is provided by Mozilla is 1.8.1.1 and the latest version is 2.2.1.
The latest version has many improvements and bug fixes from 1.8.1.1.

This issue is open for 2 years.
Unlike other bugs that users can contribute patches for, fixing this bug requires Mozilla's decision and contribution.
I hope that someone from Mozilla will be assigned to fix this bug.

(In reply to Takanori MATSUURA from comment #17)

OpenH264 plugin which is provided by Mozilla is 1.8.1.1 and the latest version is 2.2.1.

Latest is 2.2.0, not 2.2.1. I have reached out to Guangwei Wang about a 2.2.1 status but no response. Others are waiting on the recent issues (PR3477 and PR3478) to get back-ported so a properly working release can get pushed. But judging by how glacially slow it went from 2.1.x to 2.2.x, we might be talking about this 2 years from now with zero action.

Of course, nothing is stopping someone from building from source with the fixes for PR3477 and PR3478 added.

(In reply to Arthur K. [He/Him] from comment #18)

Latest is 2.2.0, not 2.2.1.

Sorry.
I meant to say 2.2.0.

(In reply to Takanori MATSUURA from comment #19)

(In reply to Arthur K. [He/Him] from comment #18)

Latest is 2.2.0, not 2.2.1.

Sorry.
I meant to say 2.2.0.

Mondainai.

Brief update, we (media team) discussed this issue today, and we've put this issue into our roadmap and would keep discuss what the priority for this issue is depending on our current resources.

One of the last times I got a response back from Guangwei Wang regarding a possible 2.2.1 release was "Got it. So I suppose that you will create other PRs for Frame ordering bug. Will public v2.2.1 after this PRs are merged."

So it seems whomever on the OpenH264 / github side that needs to merge the fixes for a 2.2.1 release needs to step up and do so and have Guangwei Wang push a release.

v2.3.0

• Decoder bug fix to correct the B-frame order in some case
• Fixes to avoid hanging for multi-thread decoding
• Some Enhancements and fixes for rate control
• Assembly optimization for loongson platform
• Some minor bug fixes

This is good news. At least we're not waiting another year+ for the bugfixes.

Is 2.3.0 now in suitable shape for inclusion into FF 106?

With FF112 on the horizon, is bug 1795014 and any remaining bugs in OpenH264 2.3.1 blocking it likely making it into 112?

(In reply to Arthur K. (he/him) from comment #27)

With FF112 on the horizon, is bug 1795014 and any remaining bugs in OpenH264 2.3.1 blocking it likely making it into 112?

We roll out our GMP plugins through a different release process, so when the new plugin ships, it will go out to existing releases.

I can't commit to a specific timeline, as there is a process we need to follow for QA testing/sign off with these off train releases. And for that, I need to request a slice of their time/resources to verify it :). But I hope in the next few months we will get a window to ship the update.

release v2.4.0

https://github.com/cisco/openh264/releases/tag/v2.4.0

Note that release 2.4.0 is broken[1]. Better wait for 2.4.1.

[1] https://github.com/cisco/openh264/issues/3711

2.4.1 is now available:

https://github.com/cisco/openh264/releases/tag/v2.4.1

• Fix off by one regression in decoder
• Add security policy
• Fix glitches that sometimes happen near the scene changes