[workspace] Rewrite metadata scraping to be less brittle #23530
Conversation
jwnimmer-tri
added
release notes: none
|
@drake-jenkins-bot linux-noble-unprovisioned-gcc-bazel-experimental-mirror-to-s3 |
|
@drake-jenkins-bot linux-noble-unprovisioned-gcc-bazel-experimental-mirror-to-s3 (I have a PR coming with the docs on this delay ... using this as an example because (a) I noticed it didn't trigger as you'd expected and (b) it's a good chance to remind myself of the language Jenkins uses on the statuses). |
|
(Ah, the "please" would help...) @drake-jenkins-bot linux-noble-unprovisioned-gcc-bazel-experimental-mirror-to-s3 please. |
|
+assignee:@Aiden2244 for feature review (or delegation), please? |
There was a problem hiding this comment.
Ran the new_release script on this branch and everything is working as expected.
+a:@sammy-tri for platform review (per schedule), please.
@Aiden2244 reviewed 73 of 73 files at r1, all commit messages.
Reviewable status: LGTM missing from assignee sammy-tri(platform)
There was a problem hiding this comment.
I need to spill this, running out of time for today. -a:@sammy-tri +a:@rpoyner-tri
Reviewable status: LGTM missing from assignee rpoyner-tri(platform)
Instead of fetching everything and then digging around in the output base to find magic-named json files, instead we list the required files explicitly as BUILD dependencies so that we'll fail-fast and so that we don't find stale json files from older revisions of Drake.
jwnimmer-tri
force-pushed
the
new-release-direct
branch
from
0ee2e6e to
682e6b3
Compare
Instead of fetching everything and then digging around in the output base to find magic-named json files, instead we list the required files explicitly as BUILD dependencies so that we'll fail-fast and so that we don't find stale json files from older revisions of Drake.
This change is