- User Since
- Sep 6 2017, 16:08 (110 w, 4 d)
Fri, Oct 18
Thanks for explaining. I think there's some minor improvements we can make later, but they don't need to block this diff specifically.
Wed, Oct 16
I don't see a good reason to block this based on the discussion in PR14941. That change is an improvement, but we'll need to backport up to it to get it.
Tue, Oct 15
I don't agree that this is easier to read, but I do agree with where this is heading and that it will integrate with ArgsManager better.
Mon, Oct 14
testplan doesn't verify that all modified defaults are still set as expected
For testing the defaults, the testplan should *not* include various arguments.
Fri, Oct 11
Thu, Oct 10
Wed, Oct 9
+1 to release notes
Tue, Oct 8
Mon, Oct 7
Good point on the composability aspect. I started to notice issues arising from the current design while working on D4213. It's more clear to me now that a composable design would make it more robust in addition to providing the caller with more customization of the build.
Fix according to feedback
Fix failing check-java tests
Prevent re-runs of setup bleeding the default value previously set on BUILD_DIR
Fri, Oct 4
Considering the error message, I would prefer that this is merged with D4206. Landing immediately after is acceptable as long as the diffs don't cross release boundaries, but why risk it in the first place?
Simplify build dir creation
Fixed a bug where non-existant build dir would fail
- Rebase on D4213 and take advantage of it to make this script more robust.
- Cleanup the repo state, especially in the case of failures.
- Reduced code ownership.
- Reduced cognitive load on merge conflicts.
Thu, Oct 3
Where did these translations come from? (should be in summary)