-
Notifications
You must be signed in to change notification settings - Fork 91
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Harvest the "without changes" runs from Taskcluster #865
Conversation
This change extends the set of valid task names to accept the "*-results-without-patch" tasks that are being added in web-platform-tests/wpt#14382 . The change also adds a "without_patch" label to these special runs to prepare for the next step: using these runs to calculate PR regressions (not implemented in this change).
Staging instance deployed by Travis CI! |
Staging instance deployed by Travis CI! |
shared/util.go
Outdated
@@ -33,6 +33,10 @@ const BetaLabel = "beta" | |||
// i.e. run from the master branch. | |||
const MasterLabel = "master" | |||
|
|||
// WithoutPatchLabel is the label for running just the affected tests on a PR | |||
// but without the patch. | |||
const WithoutPatchLabel = "without_patch" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So... I did say that eventually master
might stop implying a full run, and this is that day.
I'm wondering if labelling the runs that aren't just the affected tests as full
is a better route for our default homepage.
Alternatively (or even in conjunction with full
) we can pick a more succinct name for this label. How about we go the science route, and label them as control
? Or baseline
? normal
? prior
?
... I think baseline
might be best, since we use that term with the same semantic meaning in other contexts.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@foolip suggested base
(without_patch) and head
(with patch) offline. They work IMHO, WDYT?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Then I threw pr_base
and pr_head
into the mix. Either is fine, only possible misreading is that head
could mean HEAD, i.e. master.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Back to your first comment, Luke. I understand master
eventually won't be full runs in the future, but this PR doesn't change the current status that master runs are always full runs, right?
I like pr_base and pr_head, for the overuse of the term "head" itself.
We settled down on the names of Taskcluster tasks (already landed in WPT) and labels on wpt.fyi.
@lukebjerring PTAL again |
This change extends the set of valid task names to accept the
"*-results-without-changes" tasks that are being added in
web-platform-tests/wpt#14382 . The change also
adds a "pr_base" label to these special runs, and "pr_head" to the PR
runs with the changes to prepare for the next step: using these runs
to calculate PR regressions (not implemented in this change).
Related to #708 .