But beware bug 969988 while testing, and use something other than linux32/linux64 as the builder.
Another option would be to do something similar to how mochitests are handled (mochitests or mochitest-1, etc.) but I'm not sure that is any less messy.
Assignee: nobody → dminor
Status: NEW → ASSIGNED
Attachment #8397303 - Flags: review?(bhearsum)
Comment on attachment 8397303 [details] [diff] [review] Add jittest to try chooser. Review of attachment 8397303 [details] [diff] [review]: ----------------------------------------------------------------- I don't know this code very well, but I don't know who else to give this to...
Attachment #8397303 - Flags: review?(bhearsum) → review+
Adding another bit like http://mxr.mozilla.org/build/source/buildbotcustom/try_parser.py#16 would turn this from a miserable pain in the ass ("I chose -u jittest, but it didn't even run on Linux!") into a tolerable thing.
Following philor's suggestion to make this a tolerable pain in the ass.
Attachment #8398034 - Flags: review?(catlee)
Assuming the second patch is ok, maybe the first patch should just add a checkbox for jittests to run everything. Breaking out the jittest suites for linux as options might be more confusing than helpful.
Nope, you still need to have the explicit checkboxes for the parts, and this is (part of) why we should stop splitting things differently per-OS, per-debug/opt on a single OS, and every other which way we can think of splitting things every time we turn around: if your patch gets backed out for failing jittest-2, not jittest but jittest-2, you should not have to know that jittest+Linux will run exactly the jittest-2 that you have to run.
Thanks! Pushed buildbotcustom patch here: https://hg.mozilla.org/build/buildbotcustom/rev/e9a52c460141 and trychooser patch here: https://hg.mozilla.org/build/tools/rev/9be42dbbcb54
Working as expected.
Status: ASSIGNED → RESOLVED
Closed: 5 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.