This is a meta bug to track all the things that should be got into the next release of OpenH264 v1.8.0.

The bugs of 1479831, 1485229, 1485232, 1479669, 1479670, 1481139, 1485134, 1485208 are found in the master branch. The v1.8.0 branch (https://github.com/cisco/openh264/tree/v1.8-Firefox39) is not affected. So created another mega bug(https://bugzilla.mozilla.org/show_bug.cgi?id=1486988), and move all theses bugs to that one.

As the bug https://bugzilla.mozilla.org/show_bug.cgi?id=1481139 has been fixed, I think the gmp-openh264 v1.8.0 is ready to pickup. The code branch is v1.8-Firefox39: https://github.com/cisco/openh264/tree/v1.8-Firefox39. The current tip commit is https://github.com/cisco/openh264/commit/6fe15a6b82d492bebe388c55b7ee5131208e7334. As shown in the RELEASE note (https://github.com/cisco/openh264/blob/v1.8-Firefox39/RELEASES), this release is mostly a maintenance release. v1.8.0 ------ - Add meson build for Linux/Windows platform - Disable background detection for screen route - Add a workaround for Visual Studio 2013 C++ x64 compiler bug on AVX2. That bug will cause crash and has been fixed in Visual Studio 2014 - Change the default profile from baseline to high if user does not set it and CABAC is specified - Skip frames that are marked as IDR due to scene change and simultaneously marked as skip frame to reduce bit rate - Refine threshold calculation algorithms for rate control in lower frame rate to get better effect - Encoder return with a specific return value instead of uninitialize encoder when input resolution is invalid - Refine strategy on level change to avoid frequent IDR. Encoder will not be reset if level is changed to a smaller one - Support to set the min and max QP values on screen content mode - Fix a memory issue that may cause encoder crash when temporal layer change - Corrected some statistics information - Refine error concealment algorithms to improve user experience - Support to get information about current output picture is reference picture or not on decoder side - Bug fix for decoder when 8x8 prediction mode is enabled on the input bitstream - Enable NEON for ChromeOS devices - Support for Fuchsia operating systerm - Support for building arm64 with MSVC - Remove some warnings when building with MSVC - Fix clang compiler error when building arm assembly funtions - Bug fixes for unit test

At this point I lean towards skiping 1.8.0 release and try to update to something after 1.8.0 tracked in bug 1486988.

Release v1.8.1 is ready to pick up now, see Bug 1486988.

If 1.8.0 is not going to be landed can we close this as WON'T FIX?

Yes that sounds right. We won't fix this bug here and instead fix bug 1486988.