Releases/Firefox 3.5rc3/BuildNotes
Build Engineers
nthomas/bhearsum
release tracking bug
Signed-off Revision(s)
Build1: c597772f0c6c
Build2: a625a31a0ad1
Tags
On releases/mozilla-1.9.1:
| Build # | Tag | Changeset |
| 1 | GECKO191_20090623_RELBRANCH | c597772f0c6c |
| FIREFOX_3_5rc3_BUILD1 | c0571b2edfba | |
| 2 | FIREFOX_3_5rc3_BUILD2 FIREFOX_3_5rc3_RELEASE | a625a31a0ad1 |
Build data
| Type | Build ID | Build machine |
| [Windows installer/zip] | ||
| [Mac compressed] | ||
| [Linux compressed] |
Notes
Build 1
- cleaned up directories on slaves from 3.5rc2 build2
- patch buildbot-configs/mozilla2/release_config.py, and get review. Add the driver specified changeset and land the change
- land l10n-changesets bump from bug 498687
- reconfig production-master
- do sendchange
buildbot sendchange --username=nthomas --master=localhost:9010 --branch=releases/mozilla-1.9.1 -m "Firefox 3.5rc3 build1" goforit
Tag
No problems with tagging per se. The slave pool was very busy with other work and we had to wait a few minutes to get a machine. And it took 49 minutes instead of 12 to tag, on moz2-linux-18/Intel01 cluster vs moz2-linux-slave20/Intel02 (3.5rc2build2).
Source
No problems.
Build/Repack
No problems. Removed stage:.../3.5rc3-candidates/build2/en-US.xpi.
RESPIN DECLARED
Build 2
- cleaned up directories on slaves from 3.5rc3 build1
- patch buildbot-configs/mozilla2/release_config.py
- reconfig production-master
- do sendchange
buildbot sendchange --username=nthomas --master=localhost:9010 --branch=releases/mozilla-1.9.1 -m "Firefox 3.5rc3 build2" goforit
Tag
No problems. 12 mins elapsed on moz2-linux-slave07, slave pool less busy.
Source
No problems.
Build/Repack
No problems
Sign
instructions No problems
L10nVerify
No problems, just comment changes and a valid change in rm.
Generate updates and push betatest snippets
Like rc2, had to modify the patcher config to get correct snippets for the betas. Additionally, I modified the test snippets for rc1 and rc2 to use 'extv=3.5' while backupsnip was running (so we don't have to have a second set of snippet dirs to fix them, again). Did that with the following:
# bhearsum @ aus2-staging
# fix test snippets
cd ~/
rsync -av /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3-test/ 20090624-Firefox-3.5rc3-test/
cd 20090624-Firefox-3.5rc3-test/Firefox
rm -rf 3.1a1 3.1a2 3.1b1 3.1b2 3.1b3 3.5b4 3.5b99
find . -type f -exec sed -i -e 's/extv=3.5rc3/extv=3.5/' {} \;
# this next command should print out a bunch of 'extv=3.5' lines
find . -type f -exec grep extv {} \;
# now, beta+release channel
cd ~/
rsync -av /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3/ 20090624-Firefox-3.5rc3/
cd 20090624-Firefox-3.5rc3/Firefox
rm -rf 3.1a1 3.1a2 3.1b1 3.1b2 3.1b3 3.5b4 3.5b99
find . -type f -exec sed -i -e 's/extv=3.5rc3/extv=3.5/' {} \;
# this next command should print out a bunch of 'extv=3.5' lines
find . -type f -exec grep extv {} \;
# rsync the modified snippets incoming the staging snippet dirs # test first cd ~ # should only see 'Firefox/' rsync -n -av 20090624-Firefox-3.5rc3-test/ /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3-test/ | grep -v '\(3.1a*\|3.1b*\|3.5b*\)' rsync -n -av 20090624-Firefox-3.5rc3/ /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3/ | grep -v '\(3.1a*\|3.1b*\|3.5b*\)'
# cltbld @ aus2-staging # backup original snippets cd /opt/aus2/snippets/staging rsync -av 20090624-Firefox-3.5rc3/ 20090624-Firefox-3.5rc3-bad-rc-snippets/ rsync -av 20090624-Firefox-3.5rc3-test/ 20090624-Firefox-3.5rc3-test-bad-rc-snippets/
# push fixed snippets into the staging dirs cd ~ rsync -av ~bhearsum/20090624-Firefox-3.5rc3-test/ /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3-test/ rsync -av ~bhearsum/20090624-Firefox-3.5rc3/ /opt/aus2/snippets/staging/20090624-Firefox-3.5rc3/
Update verify
- no update from 3.5b4 to 3.5rc3 for tr and mr (known patcher bug), should fill this hole in
- can't verify updates to 3.5b4 for the same locales, as the candidates dir got removed from ftp
- the usual .chk messages on windows