Closed Bug 388586 Opened 19 years ago Closed 19 years ago

Do Firefox major update from FF1.5.0.12 -> FF2.0.0.6

Categories

(Release Engineering :: General, defect, P1)

x86
All
defect

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: moco, Assigned: rhelmer)

References

()

Details

Attachments

(2 files)

I just installed 1.5.0.12 and manually checked for update. I'm getting offered 2.0.0.4, see https://aus2.mozilla.org/update/1/Firefox/1.5.0.12/2007050813/WINNT_x86-msvc/en-US/release/update.xml We should be offering 2.0.0.5.
I thought this was on purpose, as we had a confirmed update path from 1.5.0.x --> to 2.0.0.x (1.5.0.12 --> 2.0.0.4). Otherwise, there may be some change in some other 2.0.0.x release that breaks the 1.5.0.x major update.
OS: Windows XP → All
Priority: -- → P3
There is some work that needs to be done to prepare these updates (mostly AUS configuration, but some MAR repacking as well e.g. bug 385281). We had scheduled it to be done shortly after 2.0.0.5's planned release date, but 2.0.0.5 shipped sooner than expected.
We just shipped FF2.0.0.5, so this is expected behaviour. For Q3, we will be creating a "major-update-from-1.5.0.12->2.0.0.5". Build and QA need to schedule this. In the meanwhile, users will have to update twice. This could be annoying for some users, but they will end up at the same desired end state (a secured 2.0.0.5), so are not exposed to security bugs. Keeping this bug open, as tracking bug.
Summary: 1.5.0.12 presenting major update to 2.0.0.4 (and not 2.0.0.5) → Do Major update from FF1.5.0.12 -> FF2.0.0.5
Rob / John, thanks for the info. Should this block on bug #388370: "Change 1.5->2.0 major update text to be stronger" Perhaps we should wait before flipping the major-update-from-1.5.0.12->2.0.0.5 switch until we have updated the details HTML.
Blocks: 388370
(In reply to comment #3) > In the meanwhile, users will have to update twice. This could be annoying for > some users, but they will end up at the same desired end state (a secured > 2.0.0.5), so are not exposed to security bugs. > > Keeping this bug open, as tracking bug. Upping the severity. This is not exactly correct, since users who said "later" to the 2.0.0.4 MU won't be re-prompted for 2.0.0.5. And now, as 1.5.0.12 users, they're vulnerable to public security issues in the 1.8.0-branch builds. So, we've got a set of users who are exposed to security bugs here; don't know how big a group that is (we can find that if need be), but IMO, we should schedule this ASAP after we finish the 2.0.0.5 releases, and work on getting it tested so we can offer 1-2 weeks after Firefox 2.0.0.5 went out the door.
Severity: normal → critical
1) updated subject to point to latest release on 2.x branch. 2) Is new stronger text approved? Offline discussions continue, even though bug#388370 is closed. 3) New text would need to be translated?
Summary: Do Major update from FF1.5.0.12 -> FF2.0.0.5 → Do Firefox major update from FF1.5.0.12 -> FF2.0.0.6
John, for item 2 in comment 7, what is there left here to discuss, the site is live for en-US? I've CC'ed Pascal on the bug for translation assistance/timeframe. Not sure what that entails. We should probably not gate on the translations.
(In reply to comment #7) > 3) New text would need to be translated? Pascal just pointed me to bug#390792 which is tracking the translation work.
Assignee: build → nrthomas
Priority: P3 → P2
No longer blocks: 388370
Depends on: 388370, 390792
Changes include * speculatively adding ro to the list of locales (still awaiting sign off from l10n/release-drivers for the 2.0.0.x version to go out of beta) * switch to the release version of files wherever possible, and let beta-test go to bouncer rather than stage. Snippets shouldn't differ between channels now (no need in this case), except for the repacked CJKT locales where I'm planning to follow your lead and do manually.
Attachment #278596 - Flags: review?(rhelmer)
Attachment #278596 - Flags: review?(rhelmer) → review+
I'm out for a few days, so over to Rob. Current status: * all snippets generated and on the AUS staging server (20070904-Fx-1.5.0.12-major-*) * custom mars for CJKT are at http://stage.mozilla.org/pub/mozilla.org/firefox/nightly/experimental/bug388586/ same as last time, the *-2* versions are the repacked files * update is live on betatest, QA testing going well To do: * (sanity check release notes page) * check with Pascal that we are good to go on the localized info snippets (hint! :-)) * confirm with QA & drivers that testing is complete and we should proceed * groom and push out CJKT locales to mirrors (and place dummy text file for en-US) * configure bouncer, wait 4-5 hours * QA test CJKT on releasetest * (any checks you want to do on snippets) * push update snippets live (20070904-Fx-1.5.0.12-major, or 20070904-Fx-1.5.0.12-major-beta if anyone wants to do a beta first) * ... * profit!!
Assignee: nrthomas → rhelmer
(In reply to comment #11) > To do: > * (sanity check release notes page) Sorry, this is a bit vague. I'm referring to the Build release worksheet rather than anything on mozilla.com.
Status: NEW → ASSIGNED
Priority: P2 → P1
Released yesterday.
Status: ASSIGNED → RESOLVED
Closed: 19 years ago
Resolution: --- → FIXED
Attached file rewrite_snippets.pl —
This script rewrites the CJKT snippets after the custom mar is created. There are hardcoded buildID's, paths , & URLs in it, so it would need modification if it was used again.
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: