Activity
Yesterday
dmitshur commented on golang.org/x/lint/README: make it clear that golint is deprecated and frozen15h
@pelishenkoaa@gmail.com There's no need to review old and submitted CLs, so please avoid posting empty review votes. If you'd like to contribute to Go, please see https://go.dev/doc/contribute. Thank…dmitshur reviewed +2 on cmd/go: split reuse_hg test to allow parts to run in parallel16h
Thanks.dmitshur reviewed +2 on cmd/go/internal/vcweb: hold hg server lock while serving16h
Thanks.
Similar suggestion to include a "For #nnn." line here, and the rest of the stack.dmitshur reviewed +2 on cmd/go: move go_unix_test tests to cmd/go/internal/verylongtest16h
Thanks.
(nit) https://go.dev/doc/contribute#main_content suggests "The text should be wrapped at around 72 columns." Not a big deal here since the rest of the CLs with bigger commit messages already…dmitshur commented on cmd/go: isolate some very long tests19h
(1 comment)
In case you haven't already considered it, another potential path to a similar outcome without needing to move tests between packages is to write down (and maintain) somewhere a skip lis…dmitshur reviewed +2 on cmd/go: isolate some very long tests20h
Thanks.dmitshur commented on cmd/go: isolate some very slow tests21h
(1 comment)
Just confirming that since our current test sharding implementation can only split tests at the package boundary (i.e., it's not yet able to split tests within a single large package acr…dmitshur reviewed +2 on cmd/go: isolate some very slow tests21h
Thanks.
Just to check my understanding: it's already possible to not run these tests by doing `go test -short cmd/go`, but your intention here is to make it easier to run `go test -short=false cmd/g…dmitshur commented on golang.org/x/image/webp: allow VP8L + VP8X(with alpha)22h
Thanks.This Week
dmitshur reviewed +2 on golang.org/x/tools/go/analysis/checker: suppress TestPassModule on WASI1d
Thanks.
x_tools-gotip-linux-amd64-misccompile failure ("internal compiler error: bad live variable") is unrelated to this change. It's likely already tracked in issue #78081. Bypassing since everyth…dmitshur closed an issue golang.org/x/telemetry/cmd/gotelemetry/internal/view: unrecognized failures1d
dmitshur closed an issue golang.org/x/build/buildenv: unrecognized failures1d
dmitshur closed an issue golang.org/x/tools/benchmark/parse: unrecognized failures1d
dmitshur closed an issue golang.org/x/telemetry/cmd/stacks: unrecognized failures1d
dmitshur opened a change cmd/link: use 13.0.0 OS version for macOS linking1d
Go 1.27 will require macOS 13 Ventura or later, so macOS 12 will be
unsupported.
For #75836.
For #77944.dmitshur closed an issue cmd/go: go mod init wrong version1d
dmitshur closed a change golang.org/x/build/chore: Migrate gsutil usage to gcloud storage1d
Closing since all uses of gsutil in this repository are removed as of CL 749081. Thanks.dmitshur commented on golang.org/x/build/...: migrate to 'gcloud storage' from gsutil1d
Thanks.
Now should be okay.dmitshur closed an issue cmd/go: TestScript/version_build_settings failures1d
dmitshur commented on golang.org/x/build/perf: move perf command to x/build/cmd/perf1d
(2 comments)
Done in PS 2.
I filed a separate issue to clean up the unused App Engine app.
Done.
I mentioned them both since right now https://perf.golang.org/ has no link or any indication that …dmitshur commented on cmd/link: set darwin LC_BUILD_VERSION SDK version field higher by default1d
CC @golang/darwin.dmitshur reviewed +2 on golang.org/x/build/...: set correct IAP audience1d
Thanks.dmitshur commented on golang.org/x/build/perf: move perf command to x/build/cmd/perf1d
(1 comment)
Thanks for correcting this. For some reason I was under the impression the ingress at https://cs.opensource.google/go/x/build/+/master:deploy/build-ingress.yaml;l=116;drc=7bf60f09df7b988…dmitshur commented on cmd/compile: unexplained build flakes1d
Issue #78074 is due to [CL 751260](https://go.dev/cl/751260) adding an unconditional write to `ml.data[2]`, and the `ml.data` slice being given a size of `len(load.cmd.data)` which has few than 3 ele…dmitshur commented on cmd/compile: fix arm direct tail call2d
(1 comment)
It's probably fine that p.To.Name is set to obj.NAME_NONE here and seemingly not changed to another value later, but calling it out just in case.dmitshur commented on golang.org/x/build/main.star: remove known issue for darwin-arm64_262d
Thanks.
https://ci.chromium.org/ui/p/golang/builders?q=-darwin-arm64_26 is all green except go1.25-darwin-arm64_26, but that one should be fixed by the stack ending with CL 753681.dmitshur commented on cmd/cgo/internal/test: [release-branch.go1.25] use (syntactic) constant for C array bound2d
That will get resolved in the later CL 753681. Bypassing.dmitshur commented on cmd/internal/testdir: [release-branch.go1.25] unify link command2d
All remaining trybot failures are fixed in the later CL 753681 and some unrelated timeouts. Bypassing.dmitshur opened a change cmd/internal/testdir: [release-branch.go1.25] pass -buildid to link command2d
The new dynamic loader in macOS 26 beta doesn't like binaries
without LC_UUID. Binaries built by "go build" have LC_UUID by
default. When invoking the linker manually, it has an LC_UUID by
default if…dmitshur fixed an issue cmd/internal/testdir: Test/fixedbugs/{issue42032,issue51733,issue40954}.go fail with "missing LC_UUID load command" on macOS 26 in Go 1.252d
Closing as fixed in Go 1.26 by [CL 692876](https://go.dev/cl/692876).dmitshur commented on cmd/internal/testdir: Test/fixedbugs/{issue42032,issue51733,issue40954}.go fail with "missing LC_UUID load command" on macOS 26 in Go 1.252d
Closing as fixed in Go 1.26 by [CL 692876](https://go.dev/cl/692876).dmitshur commented on cmd/internal/testdir: Test/fixedbugs/{issue42032,issue51733,issue40954}.go fail with "missing LC_UUID load command" on macOS 26 in Go 1.252d
Thanks. Those CLs only modify a package that contains tests, and isn't used otherwise, so they should be safe. It's all that's left to get Go 1.25's all.bash passing on macOS 26, so I suggest we back…dmitshur commented on cmd/internal/obj: fix indirect tail call code2d
The failure mode of linux/arm changed, but it's still not working yet. Filed #78053 for it.dmitshur opened an issue cmd/compile: linux/arm stalls during bootstrap at tip as of "use tail calls for wrappers for embedded interfaces"2d
[CL 751465](https://go.dev/cl/751465) fixed #77781 but caused linux/arm to run into a https://ci.chromium.org/b/8687988093932136273 failure mode. As of [CL 752560](https://go.dev/cl/752560), the fail…dmitshur reviewed +2 on cmd/go/internal/modfetch/codehost: cache some vcs ops3d
Thanks. Be mindful that I'm not very familiar with potential nuances here but the change seems reasonable.
In contrast to vcsRepo.Stat, where rev == "latest" is cached like any other revision, here …dmitshur reviewed +2 on cmd/go/testdata/script: rewrite mod_invalid_version to not use network3d
Thanks.dmitshur reviewed +2 on cmd/go/testdata/script: update more tests to not use network3d
Thanks.dmitshur reviewed +2 on cmd/go/internal/testdata: change list_std_vendor to use local modproxy3d
Thanks.dmitshur reviewed +2 on cmd/go/testdata/script: update more tests to not use network3d
Thanks.
(Asking to learn.) Is there some functionality that maps this .txt file's txtar content with .mod to a go.mod file? Yes, found it via [README](https://cs.opensource.google/go/go/+/master:src…Last Week
dmitshur commented on cmd/cgo/internal/test: [release-branch.go1.25] use (syntactic) constant for C array bound5d
Fixing the build error uncovers some failing tests in test/fixedbugs. Filed #78012. I tested skipping them locally, and with that all.bash passes.dmitshur opened an issue cmd/internal/testdir: Test/fixedbugs/{issue42032,issue51733,issue40954}.go fail with "missing LC_UUID load command" on macOS 26 in Go 1.255d
### Go version
```sh
# (on macOS 26)
$ go version
go version go1.25.8 darwin/arm64 # with go.dev/cl/752520 applied on top
```
### What did you do?
Ran `go test -count=1 cmd/cgo/internal/test`.
#…dmitshur closed an issue golang.org/x/build: OOM on linux-ppc64le-power10osu builder5d
The watchflakes pattern only matches the legacy builder, not the newer one that was moved to LUCI. If this happens on the new builder, watchflakes will open a new issue.dmitshur closed an issue golang.org/x/build: "unexpected stale targets" on plan9-arm5d
The watchflakes pattern only matches the legacy builder, not the newer one that was moved to LUCI. If this happens on the new builder, watchflakes will open a new issue.dmitshur closed an issue all: timeouts on `linux-s390x-ibm-race`5d
The watchflakes pattern only matches the legacy builder, not the newer one that was moved to LUCI. If this happens on the new builder, watchflakes will open a new issue.dmitshur closed an issue all: test failures with `device not configured`5d
This builder doesn't exist now.dmitshur closed an issue golang.org/x/build/internal/relui: TestRelease/rc failures5d
The relevant part of the test log is:
```
buildrelease_test.go:743: task illumos-amd64: Reproduce distpack on Windows: error: script failed: exit status 139 output:
+ set -o pipefail
…dmitshur reviewed +2 on test/fixedbugs: increase issue46234 test timeout5d
Thanks. FWIW, I found this because the [gotip-darwin-amd64_15](https://ci.chromium.org/ui/p/golang/builders/ci/gotip-darwin-amd64_15) builder being added (issue #77997) ran into this 4 times in a row…dmitshur opened a change cmd/cgo/internal/test: [release-branch.go1.25] use (syntactic) constant for C array bound6d
A test in C has an array bound defined as a "const int", which is
technically a variable. The new version of C compiler in Xcode 26
beta emits a warning "variable length array folded to constant
arra…dmitshur fixed an issue cmd/cgo/internal/test: build error on macOS 26 in Go 1.256d
dmitshur fixed an issue cmd/cgo/internal/test: build error on macOS 26 in Go 1.256d
Closing as fixed in Go 1.26 by [CL 692877](https://go.dev/cl/692877).dmitshur commented on cmd/cgo/internal/test: build error on macOS 26 in Go 1.256d
Closing as fixed in Go 1.26 by [CL 692877](https://go.dev/cl/692877).dmitshur commented on cmd/cgo/internal/test: build error on macOS 26 in Go 1.256d
Thanks. Since this is a test-only fix, and should be safe, we can probably backport it. (Another option would be to skip the test, but that doesn't seem to be needed.)
@gopherbot Please consider thi…dmitshur opened an issue cmd/cgo/internal/test: build error on macOS 26 in Go 1.256d
The recently added darwin/arm64 builder with macOS 26 (issue #76798) uncovers a failing test on [release-branch.go1.25](https://ci.chromium.org/ui/p/golang/builders/ci/go1.25-darwin-arm64_26):
I can…dmitshur commented on golang.org/x/build/main.star: add darwin-amd64_15 builder6d
Thanks.dmitshur commented on golang.org/x/build/main.star: unskip linux/mipsle builders for golang.org/x repos6d
Thanks.dmitshur opened a change golang.org/x/build/main.star: add darwin-amd64_15 builder6d
For golang/go#77997.dmitshur opened an issue golang.org/x/build: add darwin/amd64 builders with macOS 15 (macOS Sequoia)6d
This is a tracking issue for adding darwin/amd64 builders with macOS 15 (macOS Sequoia).
We already have darwin/arm64 builders ones with it as of #73397.
CC @golang/release, @golang/darwin.dmitshur opened an issue golang.org/x/build/cmd/relui: automate process of upstreaming private-track security CLs on release day (for the main branch)6d
One of the things left out of initial scope of #77214 was:
> upstreaming to main branch - this has a lot in common to release branches, but we don't have an existing ref for this and there are more …dmitshur commented on io: add an `Err` field to `LimitedReader`6d
The CL still needs to complete the [code review process](https://go.dev/doc/contribute#reviews). It already has reviewers added to it, but perhaps they haven't found the time to get to it. Since it's…dmitshur opened a change golang.org/x/build/{influx,perf}: build Docker images via Cloud Build instead of local Docker6d
For golang/go#70913.dmitshur commented on golang.org/x/build/cmd/relui: request 8 GiB of ephemeral storage6d
Thanks.dmitshur commented on net: TestListenIPv6WildcardAddr failures [consistent failure]6d
Oh, I missed the earlier https://github.com/golang/go/issues/77945#issuecomment-3999348983 where it was mentioned. That looks like an improvement as far as I can tell, so if you don't mind please do …dmitshur opened a change golang.org/x/build/cmd/relui: request 16 GiB of ephemeral-storage6d
Relui may need more scratch space to work with when running tasks
that involve cloning large repositories, such as when upstreaming
private-track security CLs on the release day.
The previous value …dmitshur reviewed +2 on golang.org/x/website/internal/history: document Go 1.26.1 and 1.25.8 releases1w
Thanks.
If this intends to refer to today's date, then:
```suggestion
Date: Date{2026, 3, 5}, Version: Version{1, 26, 1},
```
(Same below.)dmitshur reviewed +2 on net: skip TestListenIPv6WildcardAddr on DragonFly BSD and OpenBSD1w
Thanks.dmitshur commented on net: TestListenIPv6WildcardAddr failures [consistent failure]1w
The TestListenIPv6WildcardAddr test was recently added in [CL 743080](https://go.dev/cl/743080) (CC @ravisastryk, @nicholashusin), and it's consistently failing on openbsd/amd64 and openbsd/ppc64 por…dmitshur closed an issue cmd/link: unknown relocation type 7 when linking resource.syso on windows/amd64 with Go 1.26.01w
dmitshur commented on access: may-start-trybots for George Adams1w
CC @golang/release, @cherrymui.dmitshur opened a change golang.org/x/build/main.star: unskip linux/mipsle builders for golang.org/x repos1w
The builder has been added and it's passing all.bash in the main repo.
Move it from TBD_CAPACITY_HOSTS to LOW_CAPACITY_HOSTS and drop its
skip_x_repos option to get closer to fully removing its known…dmitshur reviewed +2 on image/jpeg: make decoder.receiveExtend branchless1w
Thanks.dmitshur commented on cmd/go: subcommands to add and remove modules from the module cache1w
@bysiber Please avoid posting many repetitive comments in rapid succession in the Go issue tracker. Thanks.dmitshur commented on cmd/go: update "GOPATH environment variable" section to be module-mode first1w
[CL 664235](https://go.dev/cl/664235) updated this section. @matloob Was this issue perhaps intended to be closed by it? (Its commit message said "Fixes #57355" which seems to be 404, and this issue …dmitshur commented on golang.org/x/website/_content/blog/allocation-optimizations: fix typo1w
(1 comment)
Indeed. There's no merge conflict because both CLs made an identical change here, so once rebased, only the fix on line 279 will be applied. Rebasing would be required if there was a mer…dmitshur reviewed +2 on golang.org/x/website/_content/blog/allocation-optimizations: fix typo1w
This was resolved in patch set 2.
Thanks.dmitshur commented on crypto/tls: use %w rather than %s in fmt.Errorf1w
I suggest starting with a brief issue describing this change, and include `CC @golang/security` to let the [owners](https://dev.golang.org/owners) of these packages know. They'll have a better idea w…dmitshur commented on crypto/tls: use %w rather than %s in fmt.Errorf1w
Should a change like this also have tests, updated documentation, and a reference to an issue (or a proposal if this is considered notable enough)? Possibly also a release note for Go 1.27?
Without …dmitshur reviewed +2 on golang.org/x/build/main.star: add darwin-arm64_26 builder1w
Thanks.
👍dmitshur reviewed +2 on golang.org/x/build/cmd/makemac: add darwin-arm64-26 bots1w
Thanks!dmitshur commented on golang.org/x/build/cmd/maintnerd: build Docker images via Cloud Build1w
Thanks.dmitshur commented on golang.org/x/build/maintner/maintnerd: move maintnerd command to x/build/cmd/maintnerd1w
Thanks.dmitshur reviewed +2 on golang.org/x/build/cmd/makemac: update darwin-amd64-13/141w
Thanks.dmitshur opened a change golang.org/x/build/maintner/maintnerd: move maintnerd command to x/build/cmd/maintnerd1w
Make it easier to maintain (and unify) deployment files for x/build
services by having the commands uniformly live in the cmd directory.
For golang/go#70913.dmitshur opened a change golang.org/x/build/cmd/maintnerd: build Docker images via Cloud Build1w
The intention of CL 707295 was to do this for all x/build services,
but maintnerd was missed.
For golang/go#70913.dmitshur reviewed +2 on golang.org/x/build/cmd/makemac: update darwin-amd64-121w
Thanks.dmitshur commented on golang.org/x/tools/go/analysis/passes/modernize: [internal-branch.go1.26-vendor] fix rangeint rewriting1w
(1 comment)
> Didn't know gopherbot will close issues for CLs landed on internal-vendor branch.
Ah, I was mistaken when I anticipated that happening, sorry for a misleading comment. It indeed won't…dmitshur commented on cmd/fix: rangeint uses target platform's type in the range expression, breaking other platforms1w
@dagood The change in x/tools repo in [CL 747560](https://go.dev/cl/747560) hasn't been vendored into the main Go repo yet. That will certainly happen in time for Go 1.27 (this issue's milestone), tr…dmitshur commented on golang.org/x/tools/go/analysis/passes/modernize: [internal-branch.go1.26-vendor] fix rangeint rewriting1w
(1 comment)
Note that this CL alone won't fix #77899; the CL that pulls this change into the [main Go repo's release branch](https://cs.opensource.google/go/go/+/release-branch.go1.26:src/cmd/go.mod…dmitshur commented on robpike.io/ivy/...: read version from Go metadata1w
As an alternative to running 'go build', 'go clean', then 'go install', would it work to skip the first two steps and run 'go install' right away? If there's a build error, 'go install' would report …Earlier
dmitshur commented on cmd/go: change `go mod init` default go directive back to 1.N [1.26 backport]1w
Added a release-blocker label to indicate the release shouldn't happen while this issue is still open.dmitshur commented on cmd/compile: internal compiler error: panic during rewrite tern with GOEXPERIMENT=simd [1.26 backport]1w
@dr2chase This backport issue is still open but without CherryPickApproved nor CherryPickCandidate label. Should this go back to approved but with [CL 749241](https://go.dev/cl/749241) as the new CL …dmitshur commented on cmd/fix: stringscut rewrite changes behavior [1.26 backport]1w
We reviewed and approved this during a release meeting, moved CherryPickApproved from https://github.com/golang/go/issues/77737#event-23078589596 to this issue. (CC @dr2chase.)dmitshur commented on cmd/fix: stringscut rewrite changes behavior1w
@gopherbot Please create a backport issue for Go 1.26 (original request was in https://github.com/golang/go/issues/77737#event-23012415692).dmitshur commented on cmd/fix: stringscut rewrite changes behavior1w
This issue was marked as fixed by [CL 747900](https://go.dev/cl/747900), but that CL is in the main branch of x/tools. That's sufficient to fix the problem in time for Go 1.27, but this issue is curr…dmitshur closed an issue cmd/compile: internal compiler error: 'main.func1': not lowered: v15, Load STRUCT PTR SSA [1.25 backport]2w
dmitshur commented on cmd/fix: stringsbuilder ignores variables if they are used multiple times [1.26 backport]2w
We reviewed and approved this during a release meeting, moved CherryPickApproved from https://github.com/golang/go/issues/77659#event-23078585370 to this issue.dmitshur commented on cmd/fix: stringsbuilder ignores variables if they are used multiple times2w
@gopherbot Please create a backport issue for Go 1.26 (original request was in https://github.com/golang/go/issues/77659#event-23004957221).dmitshur commented on cmd/fix: waitgroup leads to a compilation error2w
@gopherbot Please create a backport issue for Go 1.26 (original request was in https://github.com/golang/go/issues/77754#event-23010399320).dmitshur commented on cmd/fix: minmax breaks select statements2w
@gopherbot Please create a backport issue for Go 1.26 (original request was in https://github.com/golang/go/issues/77671#event-23007156867).dmitshur commented on cmd/fix: reflect.TypeOf(nil) transformed into reflect.TypeFor[untyped nil]()2w
@gopherbot Please create a backport issue for Go 1.26 (original request was in https://github.com/golang/go/issues/77636#event-23009845781).