Releases/Firefox 4.0b5/BuildNotes

From MozillaWiki
Jump to: navigation, search

Notes About Releasing

Please update the Release:Primer for future releases (bug fixes, changes to automation) as needed

Build Engineers

lsblakk

Signed-off Revision(s)

L10N changesets

l10n dashboard

Tags

Build # Tag Changeset
1 GECKO20b5_20100831_RELBRANCH b6cc0d1d0fda
FIREFOX_4_0b5_BUILD1, FIREFOX_4_0b5_RELEASE 58575263536b

Build data

Build # Type Build ID Build machine Time to build
1 Linux [compressed .tar.bz2] 20100831064831 mv-moz2-linux-ix-slave02 56 mins, 42 secs
Linux64 [compressed .tar.bz2] 20100831065840 moz2-linux64-slave02 1 hrs, 1 mins, 15 secs
Macosx [.dmg] 20100831070010 bm-xserve11 3 hrs, 33 mins, 5 secs
Macosx64 [.dmg] 20100831070403 moz2-darwin10-slave05 1 hrs, 51 mins, 7 secs
Win32 [.exe/.zip] 20100831080731 mw32-ix-slave02 2 hrs, 42 mins, 5 secs

Build 1

  • Marked all builders for all branches (picking just one branch doesn't work right now) releases on pm for a clobber -- https://build.mozilla.org/clobberer/ has a Release Clobber option if you are not connected to Build VPN
  • Moved Slaves to pm
    • from pm01
      • bm-xserve{06,11,12,16}
      • mw32-ix-slave{02,03,04}
      • moz2-darwin10-slave{05,06,07}
      • mv-moz2-linux-ix-slave{02,03}
    • from pm03
      • mv-moz2-linux-ix-slave{08,11}
      • moz2-linux64-slave{01,02,03}
      • mw32-ix-slave{14}
  • Landed configs patch
  • Updated /tools/buildbotcustom, removed /builds/buildbot/moz2-master/*.pyc, checkconfig
  • Closed tree for tagging and then started automation with:
buildbot sendchange --username=lsblakk --master=localhost:9010 --branch=mozilla-central -m "Firefox 4.0b5 build1" Go

Tag

mv-moz2-linux-ix-slave03

No problems, re-opened tree. Tagged buildbot-configs, buildbotcustom, tools with FIREFOX_4_0b5_BUILD1, FIREFOX_4_0b5_RELEASE

Source

mv-moz2-linux-ix-slave03

No problems

Build/Repack

  • Announce the build candidates dir link in the Releases wiki
  • Linux build/repack - no problems
  • Linux64 build/repack - no problems
  • Macosx64 build/repack - no problems
  • Macosx build/repack - no problems
  • Win32 build - no problems
    • id repack failed because it lost the slave so I did a force build setting branch=id, en_revision&l10n_revision=FIREFOX_4_0b5_RELEASE, and locale=id
    • id passed repack on second run

Returned slaves (all but 2 linux, and one per OS for post-signing)

  • mv-moz2-linux-ix-slave{02,03} -> pm01
  • moz2-linux64-slave{02,03} -> pm03
  • moz2-darwin10-slave{06,07} -> pm01
  • bm-xserve{11,12,16} -> pm01
  • mw32-ix-slave{02,03,04} -> pm01

Unittests

Triggered properly, no problems.

Results

  • Linux/Linux64/Macosx64/Win32{win2k3,win7}/Macosx
    • had a lot of browser_list.js fails in mochitest-other, filed bug 592279 after checking with #developers - this is new to b5, browser_list.js was not present in b4
  • Linux/Linux64
    • opengl test suite orange, but is also orange in trunk and not exposed/scraped yet tracking:bug 580405
  • Linux64
    • errors on mochitest-2 in test_suspend.html which matched bug 589442, made comment with log link in bug
    • tp4 data failed to upload - matched issue to bug 587812 and made comment with log link
  • Macosx64
    • hit the test_punycodeURI.js error from bug 561350, made note in bug
    • talos chrome errors: FAIL: Busted: tgfx FAIL: browser frozen
  • Win32{win7}
    • Rev3 WINNT 6.1 release-mozilla-central opt test direct3D - orange

Signing

en-US.xpi removed from stage bug 485860:

#ffxbld@stage
cd /pub/mozilla.org/firefox/nightly/4.0b5-candidates/build1
rm en-US.xpi

Using instructions.

L10N Verify

The output of l10n verify is horrifically long due to the switch to omni.jar. Every locale has the files from chrome/ab-CD.jar moved into omni.jar, and when when you're unpacking said jars and using diff -N you get a lot of output.

  • linux and linux64 look fine
  • win32 has differences between the three nss chk files, which also affected 4.0b4. should be fixed by bug 592457

Updates

Update Verify

Several problems.

  • Win32 wasn't able to complete due to slave disconnecting (bug 592490)
  • update verify configs weren't bumped properly, bug 592570 - we tried to update 4.0b3 with 4.0b5 but then compare it against 4.0b4
    • landed fix and reran (no clobber necessary for this, it lives in the tools checkout)
  • green on those runs

For each partial, output like

Only in source/bin: chrome
Only in source/bin/defaults: autoconfig
Only in source/bin/defaults: profile
Only in source/bin: modules
Files source/bin/omni.jar and target/bin/omni.jar differ
diff -r source/bin/removed-files target/bin/removed-files
69a70
> defaults/pref/reporter.js
544,545c545,746
< components/binary.manifest
< omni.jar
---
> chrome/en-US.jar
> chrome/browser.jar
...
> res/table-remove-row-hover.gif
Only in source/bin: res
Contents of source/bin/chrome dir only in source or target
Contents of source/bin/defaults/autoconfig dir only in source or target
Contents of source/bin/defaults/profile dir only in source or target
4080011    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/defaults/profile/chrome
Contents of source/bin/modules dir only in source or target
3228225    0 drwxr-xr-x   6 cltbld   Administrators        0 Aug  19:44 source/bin/modules/services-sync
3228227    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/modules/services-sync/base_records
3228233    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/modules/services-sync/engines
3228242    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/modules/services-sync/ext
3293791    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/modules/services-sync/type_records
3228278    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/modules/tabview
Contents of source/bin/res dir only in source or target
3883666    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/res/dtd
4145814    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/res/entityTables
4145821    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/res/fonts
4145830    0 drwxr-xr-x   2 cltbld   Administrators        0 Aug  19:44 source/bin/res/html
WARN: non-binary files found in diff
WARN: check_updates returned warning for WINNT_x86-msvc downloads/Firefox Setup 4.0 Beta 4.exe vs. downloads/Firefox Setup 4.0 Beta 5.exe: 2
with a big snip in removed-files. The complete is the same except that it doesn't have the removed-files change, and chrome isn't empty:
Contents of source/bin/chrome dir only in source or target
3162460  280 -rw-r--r--   1 cltbld   Administrators   572843 Aug  16:00 source/bin/chrome/be.jar

On linux & linux64 this appears as

Only in source/firefox/chrome: be.jar

because there's still an chrome/icons/ left.

Notes:

  • empty dirs left over from files moving into omni.jar are expected, updater isn't allowed to remove them
  • removed-files is not included in partial updates for some reason (see make_incremental_updates.py), the diff is harmless though
  • omni.jar differs on win32 - fixed by bug 592369 ?
  • complete updates are leaving behind chrome/ab-CD.jar - filed bug 592574 - not urgent to fix

Only the omni.jar issue has known consquences.

Final checks before push

cd /pub/mozilla.org/firefox/nightly/4.0b5-candidates/build1/
find . ! -user ffxbld # contrib/...
./contrib/solaris_pkgadd
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.opensolaris-i386-pkg.bz2.md5sum
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.opensolaris-sparc-pkg.bz2.md5sum
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.solaris-10-fcs-sparc-pkg.bz2.md5sum
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.solaris-10-fcs-i386-pkg.bz2
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.solaris-10-fcs-sparc-pkg.bz2
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.solaris-10-fcs-i386-pkg.bz2.md5sum
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.opensolaris-i386-pkg.bz2
./contrib/solaris_pkgadd/README.txt
./contrib/solaris_pkgadd/firefox-4.0b5.en-US.opensolaris-sparc-pkg.bz2
./contrib/solaris_tarball
./contrib/solaris_tarball/README.txt
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-sparc.complete.mar.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-i386.complete.mar
./contrib/solaris_tarball/firefox-4.0b4-4.0b5.en-US.opensolaris-i386.partial.mar
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-i386.tar.bz2.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-sparc.complete.mar
./contrib/solaris_tarball/firefox-4.0b4-4.0b5.en-US.solaris-10-fcs-i386.partial.mar
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-i386.complete.mar.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-i386.complete.mar
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-sparc.tar.bz2.md5sum
./contrib/solaris_tarball/firefox-4.0b4-4.0b5.en-US.solaris-10-fcs-sparc.partial.mar
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-sparc.complete.mar
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-i386.tar.bz2
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-sparc.tar.bz2
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-sparc.tar.bz2
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-i386.tar.bz2.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-i386.complete.mar.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-sparc.tar.bz2.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.opensolaris-sparc.complete.mar.md5sum
./contrib/solaris_tarball/firefox-4.0b5.en-US.solaris-10-fcs-i386.tar.bz2
./contrib/solaris_tarball/firefox-4.0b4-4.0b5.en-US.opensolaris-sparc.partial.mar
find . ! -group firefox
find . -type f ! -perm 644
find . -mindepth 1 -type d ! -perm 755 ! -name 'contrib*'
./contrib/solaris_pkgadd
./contrib/solaris_tarball
find . -maxdepth 1 -type d ! -perm 2775 -name 'contrib*'

Solaris builds showing up is normal.

Virus check

ffxbld@stage
cd /pub/mozilla.org/firefox/nightly/4.0b5-candidates/build1/
clamscan -r  # clamdscan wasn't connecting so I used clamscan (apparently armenzg had to ask IT to start the daemon last time around)
----------- SCAN SUMMARY -----------
Known viruses: 824293
Engine version: 0.96.1
Scanned directories: 495
Scanned files: 1182
Infected files: 0
Data scanned: 12520.71 MB
Data read: 10077.93 MB (ratio 1.24:1)
Time: 3165.145 sec (52 m 45 s)

Update Bouncer

Forced bouncer_submitter on pm - checked on webpage after successful run

Push to mirrors

Wait for email from driver before continuing (as I did not do so on this release)

# ffxbld@stage
rsync -av --exclude=*tests* --exclude=*crashreporter* --exclude=*.log \
--exclude=*.txt --exclude=*unsigned* --exclude=*update-backup* \
/pub/mozilla.org/firefox/nightly/4.0b5-candidates/build1/ \
/pub/mozilla.org/firefox/releases/4.0b5

Add index.html to cover the files before release time:

# ffbld@stage
cd /pub/mozilla.org/firefox/releases/4.0b5
curl -sL https://bugzilla.mozilla.org/attachment.cgi?id=456169 | sed 's/Beta 1/Beta 5/g' > index.html
for dir in `find . -mindepth 1 -type d`; do cp -pv index.html $dir/; done

Sent email to r-d that push to mirrors was done.

Final verify

  • green on second attempt (once there was enough uptake)
  • returned last slave (mv-moz2-linux-ix-slave08) to pm03
  • sent email when mirrors were absorbed for testing at 9:19am PDT

Push snippets to beta channel

# cltbld@aus2-staging
cd /opt/aus2/snippets/staging
~/bin/backupsnip 20100831-Firefox-4.0b5
  • Upon GO from release-drivers
~/bin/pushsnip   20100831-Firefox-4.0b5

Remove index.html files

Once we are fully announced (by the driver) we can remove the index.html files:

# ffxbld@stage
cd /pub/mozilla.org/firefox/releases/4.0b5/
find . -name index.html
find . -name index.html | xargs rm -v

Release

Emailed the following to metrics@m.c

Firefox,firefox,firefox,4.0b5,4.0
Firefox,firefox,firefox,4.0b6pre,4.0