(1 comment)
@roland@golang.org FYI, the tagging workflow is now capable of doing best-effort updates for these (i.e., using whatever is the latest released version available at the start of the work…
Thanks.
I think it might be the one! 🤞
I take it this'll need to be updated when eventually uncommenting these.
Well, now that it's covered in this test case, you can no longer change it to be …
@rfindley@google.com You seem to be most familiar with the [godev nested module](https://go-review.googlesource.com/q/project:telemetry+dir:godev) here. It has some additional golang.org/x dependency…
(1 comment)
This is working as intended after #73264. The x/tools is waiting on x/mod to be tagged first, so x/mod can't wait on x/tools's future v0.35.0 version. But it can be updated to its latest…
Thanks. There might be some follow up (see inline comments), but the changes here are good and will get it closer to working.
Note that for links to pages on the Go website itself, using a relative …
Issue #74323 might be relevant here.
It appears that the reason the package `golang.org/x/tools/gopls/internal/test/marker` tests are reported as failed is because of the data race that came up duri…
Thanks. The directories line up as expected.
FWIW, I think this could say "For" (or "Updates" if you prefer that) since this step to copy x/tools @ HEAD data into x/website's initial _content to imp…
Prior to CL 681439, the loop over module requirements in readRepo
was skipping non-golang.org/x dependencies, as well as golang.org/x
dependencies that are marked as "ignored". Those skips were repla…
(1 comment)
@daniel@binaryparadox.net Please note (for future CLs) that one of conventions for Go commit messages is to use "Fixes" to mark issues as resolved. From https://go.dev/wiki/CommitMessage…
(1 comment)
I remembered this now. The deploy does this to provide a reasonable fallback (as stated in the comment) for wiki. Not sure if you've seen it, but if not, consider if the gopls endpoint w…
Thanks. I think this is worthwhile to include for now, since it eliminates a path for the automatic deploys to go awry without us easily noticing it.
Please include `err` from os.Stat in the t.Fatal…
(1 comment)
Note that that builder running at patch set 10 did run the test (i.e., it didn't skip it), suggesting it has GOROOT/doc as expected: https://ci.chromium.org/ui/p/golang/builders/try/x_we…
(1 comment)
Can you please include a direct link to where you saw this? I only found that builder failing at earlier patch sets, for reasons unrelated to this. Does it still happen at patch set 10+?
Thanks.
s/This/The/
Please add a https://go.dev/doc/contribute#copyright comment and a blank line before the package comment.
tip.golang.org. We only provide tip. at the golang.org alias, not go.d…
The test failure is likely real - when I tried it locally at PS 6, visiting localhost:6060/go.dev/gopls/ displayed a directory listing rather than a page.
There's no policy requiring this. I wanted …
In general this looks very reasonable to me, extending the approach used for /wiki/ pages with details specific to gopls (e.g., defaulting to latest published gopls tag on go.dev; HEAD on tip.go.dev)…
Reopening to track removal of the [test skip](https://cs.opensource.google/go/go/+/master:src/cmd/compile/internal/ssa/stmtlines_test.go;l=106-108;drc=4a3cef2036097d323b6cc0bbe90fc4d8c7588660). (CC @…
(1 comment)
Please also [keep](https://go-review.googlesource.com/c/website/+/614595/13..14) the date set to today (2025-07-07), and we'll re-add the votes after they reset. Thanks.
Take a closer look at the 'Submit Requirements' panel, it should show that the CL currently has 3 out of 4 submit requirements satisfied. As soon as the 4th one is satisfied, it'll get auto-submitted…
(1 comment)
(nit) Please note that the Go commit message style uses a lower case letter after the colon, and no period at the end, because it's intended to complete the sentence "This change modifie…
The previous pattern 'internal-release-branch.go1.24.\d+' works for
minor releases as intended, but doesn't work for release candidates.
RCs can have PRIVATE-track security fixes included too (for ex…
(1 comment)
@roland@golang.org This getting pretty long, so maybe worth shortening? Dropping 'and' at least?
"go1.5rc1-go1.4.2-go1.3.2-checkpoint"
I can include some change in this CL if you prefe…
In the second half of the go.dev/s/release cycle, after entering the
release freeze and before releasing the first pre-release (RC 1), a
release branch for the next major Go release is cut.
Security…
Yes, it's generally okay to apply test fixes during the freeze, that's a large part part of what it's for. The text fix needs to be safe, and this looks safe.
In the general case, the moment a tag is published, it's not viable to modify its contents without potentially causing problems. Instead, it's possible for you to publish a newer version that correct…
The only potentially relevant recent builder change I can think of is [crrev.com/c/6641842](https://crrev.com/c/6641842) (CC @mknyszek) which rebalanced test sharding strategy for #65814. Its timing …
This test has started to fail very often on the gotip-linux-arm64-asan-clang15 linux specifically recently. As visible in its [history](https://ci.chromium.org/ui/test/golang/runtime.TestNeedmDeadloc…
Hi Julian,
Thank you for contributing to the Go project, and for your work of being a maintainer of the MIPS ports. We've discussed this request on the team and agreed to grant you this access. I've…
If more time is desired and the trade-offs are appropriate, also consider the path of reverting the change that caused this regression, reopening #52745, and trying again in a future cycle.
(2 comments)
Based on the log in the issue report, and the regexp pattern, is the missing 's' here in the commit message a typo? (Otherwise the regexp won't match.)
```suggestion
"Couldn't get NT_P…
The check fails for reasons unrelated to the change here. The module requires Go 1.20 but the check tries to test with Go 1.16 (which is before [go.dev/doc/toolchain](https://go.dev/doc/toolchain#sel…
The Go 1.25 release notes draft is complete by now, and there doesn't seem to be any remaining known TODOs to resolve. Thanks everyone.
It's okay for this to be open to track any further refinement …
Thanks for the report. So the problem happens with the same Go program in an older Windows LTSC and doesn't happen in a newer Windows LTSC. Is the crypto/tls implementation in Go doing something diff…
(3 comments)
Hmm, this is the first time I'm encountering this suggestion. I have seen many instances of short inline comments not being a complete sentence (e.g., `N int64 // bytes remaining` rathe…
Yes, I gave a heads up via chat and I saw that @adonovan@google.com gave it a thumbs up, so I think that's done (Alan, please let me know if you have questions or concerns). I'll copy from https://go…
Indeed, Auto-Submit does what it says on the tin as soon as all submit requirements are satisfied. :) Fortunately there weren't any hidden test failures so it worked out.
Thanks.
(nit) Not a big deal, but mentioning for future reference: https://www.w3.org/TR/wasm-core-2/#wasm.
Please include something to connect this CL with a tracking issue. This is [always useful…
Thanks.
Both forms are interchangeable and okay. This is described in a bit more detail at [go.dev/wiki/CommitMessage](https://go.dev/wiki/CommitMessage#:~:text=To%20link%20a,or%20vice%20versa.).
Thanks for filing the issue. Someone familiar with this package may have more specific comments, but in general please see https://go.dev/doc/contribute on how to send a code change. I added a mentio…
LUCI trybots by default run a rough equivalent of all.bash on various ports. If all.bash doesn't fail for you when you run it locally, LUCI trybots shouldn't fail either.
If you'd like to have all.b…
Hmm, can you confirm you’ve done this part?
> (Also please be sure that the darwinpkg.InstallerOptions.GOARCH value is “amd64” in your construct installer call on line 28 when you’re giving …
Consider the output from `go test` without `-json` flag in a situation where a test timeout happens while multiple subtests are still running (as happened in #74279):
```sh
$ go version
go version g…
Looking at recent build completion times for x_tools-gotip-openbsd-amd64
shows it often takes around 20 minutes, while x_tools-gotip-linux-amd64
often takes around 5 minutes. Even in the main Go repo…
Of course. In theory you can use a regular ./make.bash and create the .tar.gz archive of the result, but it might be easier to use ./make.bash with -distpack flag—just like the real release process…
There was another instance, but this time it was with a non-zero offset in the middle of a month:
```
2025/06/18 14:21:34 service/activity/github: mailLoop: walkSegMail(github.fileSeg{file:"/2025-06…