Releases/Firefox 45.0/BuildNotes
From MozillaWiki
< Releases
Contents
Fennec
beta 11 build1
-
(RelMan) Ship the l10n milestone (build1 only) - (RelMan) Submit to Ship It
- Post-release tasks
release build 1
-
(RelMan) Ship the l10n milestone (build1 only) - (RelMan) Submit to Ship It
- interrupt here.
release build 2
-
(RelMan) Ship the l10n milestone (build1 only) - (RelMan) Submit to Ship It
- interrupt here.
release build 3
-
(RelMan) Ship the l10n milestone (build1 only) - (RelMan) Submit to Ship It
- Push to mirrors
- Post-release tasks
Issues
beta 11 build1
- during the merge we accidentally uplifted beta to 46 in the geck_version.json so it picked up various changes related to android api-15 switch that are happening in aurora that eventually broke the mobile builds. The culprit patch was backed out while keeping the correct gecko_version changes here in bug 1243796 + reconfig to re-setup the right versions there. After m-a => m-b next week, we need to remember to re-apply these changes.
release build 1
- failed at repack_7/10 on android-api-9 - no space left on device while signtool, automatic retry.
- we filed bug 1253114 to increase the disk space requirement for Android repacks
- interrupted because there was a failed branding for Fennec 45, RC build is branded as Beta - tracking bug 1253173
Explanation
https://hg.mozilla.org/releases/mozilla-release/file/6f81f476fd82/mobile/android/config/mozconfigs/android-api-11/release https://hg.mozilla.org/releases/mozilla-release/file/tip/mobile/android/config/mozconfigs/android-api-11/release is incorrect
release build 2
- all good with the build but a follow-up build is on the way to address bug 1246956
release build 3
- release runner got restarted - This was a buildbot problem - a change in the SETA data which wasn't handled in the buildbot code, so jmaher backed out the SETA change. The releases have been restarted, with a tweaked list of partials for desktop
- Looking at docs, I saw the link to post-release-tasks was listed but not push to mirrors for fennec. So I ran post-release without running push to mirrors which caused bug 1255057. Have updated doc references to clarify.
Firefox
Release build 1
-
(Releng + Relman) - prior to shipit, determine if there is a WhatsNewPage for the release and comment/uncomment http://hg.mozilla.org/build/buildbot-configs/file/bc26f621daa1/mozilla/release-firefox-mozilla-release.py.template#l38 - (RelMan) Ship the l10n milestone (build1 only)
- (RelMan) Submit to Ship It
- Set up whatsnew page
-
Publish RC to Beta channel - interrupt here.
Release build 2
-
(Releng + Relman) - prior to shipit, determine if there is a WhatsNewPage for the release and comment/uncomment http://hg.mozilla.org/build/buildbot-configs/file/bc26f621daa1/mozilla/release-firefox-mozilla-release.py.template#l38 - (RelMan) Ship the l10n milestone (build1 only)
- (RelMan) Submit to Ship It
- Set up whatsnew page
- Publish RC to Beta channel
- Push to mirrors
- Apply Throttling
- Verify throttling
- Publish in Balrog
- Post-release tasks
Issues
build 1
- one repack timeout and the usual GTK3 update_verify_beta errors
- one mac update verify failure due to complete mar d/l failure, rerun was OK
- beta_final_verification didn't run because the uptake check was blocked trying to find partials. This is the same as we had for 44.0.1. Unblocked by dropping files on archive.m.o
pub/firefox/candidates/45.0-candidates/build1/update/win64/zh-TW/firefox-39.0-45.0.partial.mar pub/firefox/candidates/45.0-candidates/build1/update/win64/zh-TW/firefox-41.0.2-45.0.partial.mar
and forcing the start_uptake_check builder again. We didn't get the 'ready to realease on beta' email because of the gtk3 update failures
- a follow-up RC is on the way to address bug 1252168 and bug 1246956
build 2
- release runner got restarted - This was a buildbot problem - a change in the SETA data which wasn't handled in the buildbot code, so jmaher backed out the SETA change. The releases have been restarted, with a tweaked list of partials for desktop
- network blip, automatic retry on a linux64 builder
- ususal suspects in terms of GTK3 update verify on beta channel