Closed Bug 649565 Opened 14 years ago Closed 12 years ago

Self-serve build API should allow starting a test build that didn't run originally

Categories

(Release Engineering :: General, enhancement, P5)

enhancement

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 793989

People

(Reporter: heycam, Unassigned)

References

Details

(Keywords: buildapi, Whiteboard: [tryserver][database])

If I push to try and specify `-u none`, and realise after the build is complete that I'd like to run say reftests on that build, it should be possible for me to do this from within the self-serve build API.
I'm going to add this to try_enhancements so that it can be investigated as such.  This is more involved than it may sound though since we would need to store the information of where the zip and test files are located (for the 10 days a try build lives) in the db in order to access them at a later time to do the sendchange to our test builders.
Severity: normal → enhancement
Priority: -- → P5
Whiteboard: [tryserver][database][self-serve]
Not sure if this should be filed as a new bug, but I would like to see a similar ability on m-i and m-c test runs. Specifically, I would like to be able to trigger a test that got coalesced and was never run on a particular changeset.
You already have that, just not directly through tbpl - open the "Self-serve Build API" link that appears to the right when you hover a push, and coalesced jobs will be listed there, confusing you since you know they didn't run on that push, but being helpful since when you retrigger them, they *will* run on that push.
So it does. Apparently I didn't see the test I was looking for the first time I looked, I must be going blind...
Keywords: buildapi
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → DUPLICATE
Whiteboard: [tryserver][database][self-serve] → [tryserver][database]
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.