Closed Bug 1174962 Opened 10 years ago Closed 10 years ago

esr38 and release don't agree on what FIREFOX_38_0_RELEASE is

Categories

(Release Engineering :: Release Automation, defect)

defect
Not set
normal

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: glandium, Unassigned)

Details

https://hg.mozilla.org/releases/mozilla-esr38/rev/FIREFOX_38_0_RELEASE says 504ec068cc33, https://hg.mozilla.org/releases/mozilla-release/rev/FIREFOX_38_0_RELEASE says 4c4dc6640c7e and there are content differences between the two This seems to happen regularly for new ESRs. Can we also sort out whatever creates those tags?
This feels like a release automation issue.
Component: Mercurial: hg.mozilla.org → Release Automation
Product: Developer Services → Release Engineering
QA Contact: hwine → bhearsum
What is the utility of comparing those two revisions ? I was expecting the comparison to be https://hg.mozilla.org/releases/mozilla-esr38/rev/FIREFOX_38_0esr_RELEASE, aka aa5e4bf7a93c https://hg.mozilla.org/releases/mozilla-release/rev/FIREFOX_38_0_RELEASE, aka 4c4dc6640c7e There are unavoidable differences between those, ie browser/confvars.sh. Lets take a look at your case though ... (In reply to Mike Hommey [:glandium] from comment #0) > https://hg.mozilla.org/releases/mozilla-esr38/rev/FIREFOX_38_0_RELEASE says > 504ec068cc33, That rev is also FIREFOX_38_0_BUILD2. > https://hg.mozilla.org/releases/mozilla-release/rev/FIREFOX_38_0_RELEASE > says 4c4dc6640c7e This is also FIREFOX_38_0_BUILD3, the final build for 38.0. So there was a merge from m-r to m-esr38 after 38.0 build2 (4th May), then 38.0esr build1 was produced on 5th May, but we needed another fix for Loop on 8th May. There was no merge to esr38 or further build, that's the issue ? > and there are content differences between the two This is really for RelMan to comment on. Some of it is Loop being disabled on esr38, so some fixes for that were not landed on m-esr38.
(In reply to Nick Thomas [:nthomas] from comment #2) > (In reply to Mike Hommey [:glandium] from comment #0) > > https://hg.mozilla.org/releases/mozilla-esr38/rev/FIREFOX_38_0_RELEASE says > > 504ec068cc33, > > That rev is also FIREFOX_38_0_BUILD2. > > > https://hg.mozilla.org/releases/mozilla-release/rev/FIREFOX_38_0_RELEASE > > says 4c4dc6640c7e > > This is also FIREFOX_38_0_BUILD3, the final build for 38.0. > > So there was a merge from m-r to m-esr38 after 38.0 build2 (4th May), then > 38.0esr build1 was produced on 5th May, but we needed another fix for Loop > on 8th May. There was no merge to esr38 or further build, that's the issue ? It would appear it is. So in fact, it's yet another side effect of those _RELEASE tags being updated for each _BUILDn. I just wish we stopped doing that.
I concur with comment #3: making tags be immutable is highly preferred. Also, this is your friendly reminder that tags by themselves are just labels and can be moved. Generally, prefer full 40 character hex SHA-1 to tags/branches since the latter can change over time.
In the release promotion world tagging will be done as a part of post-release. I'd even vote to skip it all together. :)
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.