1612722 - Tracking bug for build and release of SeaMonkey 2.53.1 Final

Status: RESOLVED FIXED

(SeaMonkey :: Release Engineering, enhancement, P1)

Description

[Frank-Rainer Grahl (:frg)]

+++ This bug was initially created as a clone of Bug #1584803 +++

This is a tracking bug for Build and Release of SeaMonkey 2.53.1 Final

Comment 1

[Frank-Rainer Grahl (:frg)]

Attachment: 1612722-versiondisplay-2531.patch

Final version patch

Comment 2

[Ian Neal]

Comment on attachment 9126910 [details] [diff] [review]
1612722-versiondisplay-2531.patch

[Triage Comment]
r/a=me

Comment 3

[Dmitry Butskoy (:buc)]

Does bug #1242294 planned to include in 2.53.1 ? (https://bugzilla.mozilla.org/show_bug.cgi?id=1242294#c28)

Comment 4

[Frank-Rainer Grahl (:frg)]

Attachment: mozconfigs-win2531b1.zip

mozconfigs and build scripts for Windows.

Comment 5

[Frank-Rainer Grahl (:frg)]

Does bug 1242294 planned to include in 2.53.1

No the patch needs more work so that Lightning gets disabled too and some testing. Maybe for 2.53.3

Comment 6

[Dmitry Butskoy (:buc)]

(In reply to Frank-Rainer Grahl (:frg) from comment #5)

Does bug 1242294 planned to include in 2.53.1

No the patch needs more work so that Lightning gets disabled too and some testing. Maybe for 2.53.3

What would be the best way:

• Disable Lightning advertising anyway when "exact FF" is set;
• Just drop Lightning advertising for SM at all (not touching Lightning itself, just the correspond SM code)?

Then I can provide some work there.

Comment 7

[Frank-Rainer Grahl (:frg)]

Attachment: mozconfigs-windows-2531.zip

wrong config

Comment 8

[Frank-Rainer Grahl (:frg)]

Lets discuss the UA change in Bug 1242294

Comment 9

[Ian Neal]

Attachment: mozconfig-2531final.tar

mozconfig files for Linux x64, i686 and mac cross compile
These were used with devtoolset-6 on CentOS 7

Comment 10

[Dmitry Butskoy (:buc)]

The size of the beta1 source at http://archive.mozilla.org/pub/seamonkey/releases/2.53.1b1/source is 277M.

The size of the final source at http://archive.mozilla.org/pub/seamonkey/releases/2.53.1/source/ is shown as 755M .

Something goes wrong?

Comment 11

[Ian Neal]

(In reply to Dmitry Butskoy from comment #10)

The size of the beta1 source at http://archive.mozilla.org/pub/seamonkey/releases/2.53.1b1/source is 277M.

The size of the final source at http://archive.mozilla.org/pub/seamonkey/releases/2.53.1/source/ is shown as 755M .

Something goes wrong?

See Bug 1618852

Comment 12

[Dmitry Butskoy (:buc)]

Attachment: seamonkey-2.53.1-mozilla-1533969.patch

Backport bug #1533969 to compile under Linux with newest glibc versions (gettid() issue).

Comment 13

[Dmitry Butskoy (:buc)]

Attachment: firefox-60-mozilla-1516803.patch

Backport bug #1516803 to compile by gcc with lto (ie. "gcc -flto'0.

Comment 14

[Dmitry Butskoy (:buc)]

"gcc -flto"

Comment 15

[Frank-Rainer Grahl (:frg)]

The glibc patch is already in the 2.53.2 patch queue since a few weeks. Need to check the other.

Comment 16

[Dmitry Butskoy (:buc)]

Attachment: seamonkey-2.53.1-mozilla-revert-1332139.patch

Current mozilla and SM code cannot be compiled with system libvpx >= 1.8.0 .
(FF's bundled is still 1.7.0, SM is 1.6.1).

There is a "frame_flags" statement in webrtc code:
https://hg.mozilla.org/releases/mozilla-release/file/tip/media/webrtc/trunk/webrtc/modules/video_coding/codecs/vp9/vp9_impl.cc#l869
which is gone now in libvpx upstream by:
https://chromium.googlesource.com/webm/libvpx/+/a66da313805f41076fd05358a65c1084f646f348

In 2.49.5, the correspond code parts was ifdefed by LIBVPX_SVC. This ifdefs was removed in bug #1332139 (in assumption no more neeeded since 1.6), but then upstream change things.
That way, 2.49.5 compiles fine with system libvpx-1.8, whereas 2.53 is not.

The patch is a backport of the bug #1332139 changes. Could be useful for usage at downstream.

Comment 17

[Frank-Rainer Grahl (:frg)]

We are done. See you with 2.53.2

Comment 18

[Frank-Rainer Grahl (:frg)]

Comment on attachment 9130026 [details] [diff] [review]
seamonkey-2.53.1-mozilla-1533969.patch

The profiled gettid() problem will be fixed in 2.53.2 by backporting the Firefox 68 and 69 patches.