Closed Bug 757829 Opened 8 years ago Closed 8 years ago

UrlPoller doesn't poll win32_signing_buildN.log for unsigned Thunderbird releases

Categories

(Release Engineering :: Release Automation: Other, defect, P2)

x86_64
Linux
defect

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: rail, Assigned: rail)

References

Details

(Whiteboard: [releases][automation])

Attachments

(3 files, 2 obsolete files)

... because of "if releaseConfig['productName'] == 'firefox':"
Attached patch get rid of productName (obsolete) — Splinter Review
* the idea is to use some patterns, like platform names instead of productName
* $ grep -P '(fennec|firefox|thunderbird)' process/release.py |wc -l
  0
* one of the hunks is more readable if applied and git/hg diffed with -b parameter
* to be tested in staging
* configs diff incoming
Attachment #626493 - Flags: feedback?(bhearsum)
Attachment #626493 - Flags: feedback?(aki)
Attached patch configs (obsolete) — Splinter Review
Comment on attachment 626493 [details] [diff] [review]
get rid of productName

I can see the hasAndroidPlatform stuff becoming ugly if we start building non-Android Fennec platforms, but since I don't know what form that will take, we probably shouldn't optimize for that now.

I wish there were a simple way to handle this via config rather than hardcoding platform behavior, but I don't think this is worse than keying off the product name.
Attachment #626493 - Flags: feedback?(aki) → feedback+
(In reply to Aki Sasaki [:aki] from comment #3)
> I wish there were a simple way to handle this via config rather than
> hardcoding platform behavior,

I thought about using something like releaseConfig['workflowType'] = 'firefox'/'fennec', but it also looks ugly :/
Attachment #626493 - Flags: feedback?(bhearsum) → feedback+
Attachment #627263 - Flags: review?(bhearsum)
Attached patch configsSplinter Review
Attachment #627264 - Flags: review?(bhearsum)
Attachment #626494 - Attachment is obsolete: true
I tested the patches in staging. So far they look good. I'm still going to run a couple of releases to be sure before I land them.
Attachment #626493 - Attachment is obsolete: true
Attachment #627265 - Flags: review?
Attachment #627265 - Flags: review? → review?(bhearsum)
Blocks: 730328
(In reply to Rail Aliiev [:rail] from comment #7)
> I tested the patches in staging. So far they look good. I'm still going to
> run a couple of releases to be sure before I land them.

Overnight release builds worked fine for Firefox and Fennec. I think we're safe to proceed with the patches if the approach looks OK to you.
Blocks: 744602
Attachment #627263 - Flags: review?(bhearsum) → review+
Attachment #627264 - Flags: review?(bhearsum) → review+
Attachment #627265 - Flags: review?(bhearsum) → review+
Comment on attachment 627263 [details] [diff] [review]
tools: use manuallySignedPlatforms

http://hg.mozilla.org/build/tools/rev/abb6ddb7ab70
Attachment #627263 - Flags: checked-in+
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
I removed --create-contri-dirs for some unknown reason...
r=jhopkins on IRC
http://hg.mozilla.org/build/buildbotcustom/rev/6995edbed36d
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.