Activity
Yesterday
dmitshur reviewed +2 on golang.org/x/build/internal/task: use relmeta definitions7h
Thanks.
Since there's now a named constant, consider using it here:
```suggestion
if patch.Track == relmeta.Public {
```dmitshur reviewed +2 on golang.org/x/build/relmeta: add ReleaseMilestone and SecurityPatch definitions7h
Thanks. Two questions about fields that stood out to me, but I expect you're more familiar with what they're intended to be.
Is `ID` right here? I expect it might've intended to be BuganizerID / `bu…dmitshur commented on golang.org/x/build/go.mod: update various dependencies7h
Thanks.dmitshur commented on golang.org/x/vulndb/...: lower go directive to 1.25.07h
Thanks for checking.dmitshur commented on golang.org/x/build/relmeta: add ReleaseMilestone and SecurityPatch definitions7h
(2 comments)
Filed issue #79001 for it.
These updates are reasonable but a slight preference to break them out into a separate CL, since it's pretty unrelated. Sent you CL 771740.dmitshur opened a change golang.org/x/build/go.mod: update various dependencies7h
golang.org/x/vulndb happens to use newer versions of some of the same
dependencies that golang.org/x/build has. Update to the same versions.
[git-generate]
go get golang.org/x/vulndb
go get go@1.25.…dmitshur commented on golang.org/x/vulndb/...: lower go directive to 1.25.08h
Thanks.dmitshur reviewed +2 on golang.org/x/build/internal/task: make workflow explicitly pick up non-PUBLIC changes8h
Thanks.
This is a part of some work tracked in one of the already open issues, right? Please add a `For golang/go#nnn.` line here.dmitshur commented on golang.org/x/build/main.star: remove Go 1.25 builders for x/vulndb8h
Thanks.dmitshur reviewed +2 on golang.org/x/build/internal/gophers: update Rob's Gerrit email9h
Thanks.dmitshur opened a change golang.org/x/build/main.star: remove Go 1.25 builders for x/vulndb9h
The Go 1.25 builders that test with GOTOOLCHAIN=local produce no useful
signal since the x/vulndb module had 1.26.0 become its minimum required
language version in CL 747200. Stop these builders for …dmitshur commented on golang.org/x/vulndb/...: lower go directive to 1.25.09h
Thanks. If that govulncheck test is the main reason this module ended up requiring as high of a language version as 1.26.2, it seems worthwhile to factor that out. Opened a tracking issue #79001 for …dmitshur opened an issue golang.org/x/vulndb: lower go directive to 1.25.0, update govulncheck test9h
The golang.org/x/vulndb module has some packages that we want to import from other golang.org/x modules, but that would propagate its current go directive 1.26.2, which is quite high. It'd be good to…dmitshur commented on golang.org/x/vulndb/...: update go version and dependencies10h
(1 comment)
Sent CL 771620 to make sure tests still pass with 1.26.2+/tip.dmitshur opened a change golang.org/x/vulndb/...: lower go directive to 1.25.010h
There's nothing in this module that requires 1.26.2 language semantics.
Lower its requirement to go 1.25.0 to make it possible for other modules
to depend on this module without in turn raising their…dmitshur commented on golang.org/x/vulndb/...: update go version and dependencies11h
(1 comment)
It seems this CL updates the Go toolchain version used during the build and deploy (in cmd/worker/Dockerfile and deploy/build.yaml), but also raises the minimum go language version for t…dmitshur commented on golang.org/x/build/relmeta: add ReleaseMilestone and SecurityPatch definitions11h
(1 comment)
It's because a requirement on the golang.org/x/vulndb module is added, and that module requires 1.26.2:
https://cs.opensource.google/go/x/vulndb/+/master:go.mod;l=3;drc=5571712294b453fa…dmitshur commented on golang.org/x/build/relmeta: add ReleaseMilestone and SecurityPatch definitions11h
(1 comment)
Is raising the x/build module minimum requirement from 1.25.0 to 1.26.2 intentional?dmitshur commented on all: fix a lot of spelling mistakes11h
(1 comment)
It seems both of these are considered valid alternative spellings of the same word, and we generally want to avoid changing spelling if it's not a fix or a clear improvement. But in this…This Week
dmitshur reviewed +2 on cmd/go: set a http user agent1d
Thanks. This seems okay and safe to me, but it would be good if you give some time before submitting so @matloob@golang.org has a chance to look and comment if there are concerns.
The rest of the te…dmitshur fixed an issue all: end support for macOS 12 in Go 1.271d
Those tasks are done now, closing again.dmitshur commented on all: end support for macOS 12 in Go 1.271d
Those tasks are done now, closing again.dmitshur commented on go.dev/wiki/MinimumRequirements: document that Go 1.27 requires macOS 131d
Thanks.Last Week
dmitshur starred honnef.co/go/stuff3d
dmitshur commented on cmd/go: set a unique user agent4d
That is a reasonable extension to consider. After all, the same reasoning can be applied to any VCS the go command invokes (git, hg, svn, and so on): having a go command-specific user agent helps wit…dmitshur commented on golang.org/x/build/cmd/gitmirror: stop serving-only instances4d
Thanks.dmitshur opened a change golang.org/x/build/cmd/gitmirror: stop serving-only instances5d
We continue to use gitmirror for its primary mirroring functionality,
but its secondary serving functionality has not been used in a while.
With LUCI builders, we're relying on having them fetch dire…dmitshur closed an issue golang.org/x/build/cmd/gopherbot: reply to commit comments saying we do not monitor them5d
dmitshur closed an issue golang.org/x/build/cmd/gopherbot: reply to commit comments saying we do not monitor them5d
dmitshur commented on golang.org/x/build/cmd/gopherbot: reply to commit comments saying we do not monitor them5d
Merging this into proposal #25450.dmitshur closed a change golang.org/x/build/commitcommentshook: politely respond to commit comments on golang/go5d
Thanks for working on this. Closing since issue #25450 is resolved.dmitshur fixed an issue github: lock conversations on GitHub commits5d
In 2026, GitHub added support to disable comments on individual commits via the UI:
https://github.blog/changelog/2026-03-25-disable-comments-on-individual-commits/
So we're able to accomplish this…dmitshur commented on github: lock conversations on GitHub commits5d
In 2026, GitHub added support to disable comments on individual commits via the UI:
https://github.blog/changelog/2026-03-25-disable-comments-on-individual-commits/
So we're able to accomplish this…dmitshur commented on cmd/go: set a http user agent5d
The vcweb/vctest package sets up an HTTP+HTTPS server for go command testing needs (e.g., [here](https://cs.opensource.google/go/go/+/go1.26.2:src/cmd/go/internal/vcweb/vcstest/vcstest.go;l=64) and […dmitshur commented on cmd/go: set a http user agent6d
This probably needs a test, right? Otherwise this change might accidentally regress during a refactor without any test catching it.
```suggestion
go.dev/issue/35699 was a declined proposal to add th…dmitshur commented on cmd/go: set a unique user agent6d
The default user agent in the net/http package also deliberately doesn't include the toolchain version. E.g., see comment [here](https://cs.opensource.google/go/go/+/master:src/net/http/request.go;l=…dmitshur closed an issue runtime:gccheckmark: TestNotInGoMetric/CgoCallAndCallback failures6d
dmitshur commented on runtime:gccheckmark: TestNotInGoMetric/CgoCallAndCallback failures6d
The :gccheckmark test variant may not be relevant to this failure, so merging into the top-level tracking issue for this test, issue #78877.dmitshur commented on net/http: new HTTP client package6d
CC @neild.dmitshur commented on runtime: setevent failed; errno=6, fatal error: runtime.semawakeup6d
Can you also try this at tip (1.27 to be) with its new GOEXPERIMENT= goleakprofile feature (see issue #75280)? Thanks.dmitshur commented on runtime: setevent failed; errno=6, fatal error: runtime.semawakeup6d
CC @golang/windows, @golang/runtime.dmitshur commented on runtime: make mutexActiveSpinSize configurable6d
CC @golang/runtime.dmitshur commented on runtime: `SIGSEGV` with a c-shared library and called from C#6d
CC @golang/runtime.dmitshur commented on cmd/go: when running 'go doc', GOPATH/src/<stdpkg> shadows stdlib when runtime.GOROOT() is empty6d
Thanks for the report. Would you like to send a CL? https://go.dev/doc/contributedmitshur commented on cmd/cgo/internal/testsanitizers: TestASAN/asan5_fail failures6d
CC @golang/compiler.dmitshur commented on cmd/cgo/internal/testsanitizers: TestASAN/asan5_fail failures6d
From issue triage, let's wait a bit to see if this happens more.dmitshur commented on runtime: map with 8 elements unnecessarily grows on existing key update6d
Thanks for filing an issue for this. If you have a fix in mind, would you like to send a CL? https://go.dev/doc/contribute.
CC @golang/runtime.dmitshur commented on cmd/go: go-import meta tag can redirect to insecure location6d
CC @golang/command-line.dmitshur commented on go.dev/wiki/PortingPolicy: make port-maintainers links more direct6d
Thanks.dmitshur reviewed +2 on go.dev/wiki: Add f4 editor with Go syntax highlighting support6d
Thanks.dmitshur reviewed +2 on go.dev/wiki/Range: wrap reference table to prevent Safari overflow6d
Thanks.dmitshur opened a change go.dev/wiki/MinimumRequirements: document that Go 1.27 requires macOS 136d
For golang/go#75836.dmitshur commented on crypto/tls: generate test certificates6d
(1 comment)
As written, this'll be willing to run any test whose name happens to contain the substring TestGenerateCertificates, but I think the intention is to run a single specific test (below). C…dmitshur commented on runtime: on 32bits arches timespec (64) definition is wrong6d
@Jorropo Can you please clarify why you requested this backport to Go 1.26 and not 1.25? Is this a regression from 1.25 or is there another reason? Thanks.dmitshur reviewed +2 on golang.org/x/build/internal/relui: update TestRelease to use supported Go versions1w
Thanks.
It matches what I found too, thanks.dmitshur commented on golang.org/x/build: add LUCI windows/arm64 builder1w
I see it connected at https://chromium-swarm.appspot.com/botlist?f=cipd_platform%3Awindows-arm64&f=pool%3Aluci.golang.shared-workers and it's started its first build at https://ci.chromium.org/b/8683…dmitshur commented on golang.org/x/build: add LUCI windows/arm64 builder1w
Here's the certificate for it: [windows-arm64-azure-qmuntal-1776787311.cert.txt](https://github.com/user-attachments/files/26939453/windows-arm64-azure-qmuntal-1776787311.cert.txt).dmitshur commented on golang.org/x/build: add LUCI windows/arm64 builder1w
Okay, thanks for checking. I've updated our internal config ([crrev.com/i/9223085](https://crrev.com/i/9223085)) to drop the "--01" suffix from the bot matching, can you try again now?
Either it sho…dmitshur commented on math/big: correct panic message caused by copy-paste1w
The absent commit message body and issue reference are okay in this case. Done.dmitshur commented on golang.org/x/build/cmd/pubsubhelper: upgrade to checking X-Hub-Signature-256 header1w
Thanks.dmitshur opened a change golang.org/x/build/cmd/pubsubhelper: upgrade to checking X-Hub-Signature-256 header1w
Our webhook deliveries include both the old and new headers, such as:
X-Hub-Signature: sha1=03f1071c7a4d6220619359a56d84bcfd50af6fde
X-Hub-Signature-256: sha256=bb7f253fda044aef63f82ed4a024170e85183…dmitshur commented on golang.org/x/build/cmd/bootstrapswarm: include response body in error message1w
Thanks.dmitshur commented on golang.org/x/build: add LUCI windows/arm64 builder1w
That internal config change landed only an hour ago, so please try again now, it should have be propagated by now. I also sent [go.dev/cl/768980](https://go.dev/cl/768980) to make it easier to see wh…dmitshur opened a change golang.org/x/build/cmd/bootstrapswarm: include response body in error message1w
If the {swarming host}/bot_code endpoint returns with something other
than 200 OK, it may help with debugging to see what the response body
was. This can be done manually out of band, but might as we…Earlier
dmitshur commented on cmd/go/internal/vcweb/vcstest: TestScripts/overview failures1w
The problem here was that two tests didn't complete after running for 36 minutes for some reason.
```
panic: test timed out after 36m0s
running tests:
TestScripts/git/vgotest1.txt (35m59s)
Test…dmitshur commented on golang.org/x/build/internal/task: fix 'gcloud storage' migration in publishPackageExtension1w
Thanks.dmitshur commented on golang.org/x/build: add LUCI windows/arm64 builder1w
I tried a bit to look into whether including the "--" in the CSR name is supported, and so far I didn't see a reason it couldn't work, I think we'll actually need to try it to be sure. So let's go ah…dmitshur opened a change golang.org/x/build/cmd/gorebuild: log at start of GerritTarGz1w
GerritTarGz prints about problems it encounters and how it handles them,
but those messages are hard to make sense of without some context about
what work was being done. Add a brief log line to help…dmitshur commented on golang.org/x/build/main.star: remove known issue for linux-arm64_debian131w
Thanks.dmitshur commented on : [release-branch.go1.26] fix incorrect loop trip counts1w
go1.26-linux-arm64_debian13 trybot failure is #78406, fixed only after this CL was prepared and unrelated to the change here, so bypassing since everything else passed.dmitshur commented on all: [release-branch.go1.26] update x/net to 705de46f1w
The change to src/net/http/h2_bundle.go seems missing (as caught by `go test cmd/internal/moddeps`). Otherwise this looks good.dmitshur opened a change golang.org/x/build/main.star: remove known issue for linux-arm64_debian131w
go.dev/issue/78405 and go.dev/issue/78406 were resolved,
and the builder has started working on release branches:
https://ci.chromium.org/ui/p/golang/builders/luci.golang.ci/go1.26-linux-arm64_debia…dmitshur commented on cmd/link: stop requiring gold on arm64 when GNU ld is fixed [1.26 backport]1w
> fixed by CL 740480 with a potential followup in CL 747660
@cherrymui and I discussed whether to wait for the potential followup. Since the original fix is enough for the debian13 builders, we'll s…dmitshur commented on cmd/link: use bfd ld 2.41+ on linux/arm64 instead of gold1w
(1 comment)
@dimitri.ledkov@surgut.co.uk Can you apply this change? Or perhaps I can edit it for you, though then I can't review this CL as I'll become the uploader.dmitshur commented on golang.org/x/build/main.star: move known_issue computation from define_builder to enabled1w
(1 comment)
Actually done this time...dmitshur commented on golang.org/x/build/main.star: move known_issue computation from define_builder to enabled1w
Thanks.
Oh, this is one more place where the earlier computation should be reused. No effect on generated files now, but it might make a different in the future. Done in PS 2.dmitshur commented on golang.org/x/build/internal/relui: update TestRelease to use supported Go versions1w
(3 comments)
Hmm, unfortunately I think having a test for 1.27 rc intersects poorly with the task to generate a x/tools CL, because that ends up trying to use the real `golang.org/dl/go1.27rc1` comm…dmitshur commented on cmd/compile: [release-branch.go1.25] handle min integer step in loop1w
The linux-arm64_debian13 builder doesn't work on release branches yet, and it currently has a known issue associated with it:
https://ci.chromium.org/p/golang/g/go-go1.25-known_issue/console
It sho…dmitshur opened a change golang.org/x/build/main.star: move known_issue computation from define_builder to enabled1w
In general, we always first call enabled on a given builder type to tell
whether a builder exists at all, and only then do we call define_builder
for that builder type.
Computing the known issue in …dmitshur fixed an issue golang.org/x/build/cmd/gorebuild: may fail to produce a useful report due to running into '429 Too Many Requests' unauthenticated rate limiting1w
https://go.dev/rebuild has a successful report with the new version ([CL 767880](https://go.dev/cl/767880)).
<img width="547" height="405" alt="Image" src="https://github.com/user-attachments/assets…dmitshur commented on golang.org/x/build/cmd/gorebuild: may fail to produce a useful report due to running into '429 Too Many Requests' unauthenticated rate limiting1w
https://go.dev/rebuild has a successful report with the new version ([CL 767880](https://go.dev/cl/767880)).
<img width="547" height="405" alt="Image" src="https://github.com/user-attachments/assets…dmitshur commented on golang.org/x/build: add LUCI aix-ppc64 builder1w
That message does seem unexpected. I also see it in a recent task like https://chromium-swarm.appspot.com/task?id=77aef1f18b21ad10 and it's blocking further progress on the task.
If you look at http…dmitshur opened a change go.dev/wiki/PortingPolicy: make port-maintainers links more direct1w
Link directly to the 'Teams' tab instead of the default 'Members' tab.
It's more useful, since the top-level golang/port-maintainers team is
not really used; it just holds the port-specific sub-teams…dmitshur commented on cmd/go/internal/modfetch: update zip hashes for new flate compression1w
Thanks.
Noting that a number of watchflakes-opened issues can be closed when this CL lands: https://go.dev/issues?q=is%3Aissue%20state%3Aopen%20author%3Agopherbot%20%22(but%20content%20has%20correct…dmitshur reviewed +2 on cmd/go/internal/modcmd: omit null fields1w
Thanks.dmitshur reviewed +2 on cmd/go/internal/workcmd: include toolchain and godebug in -json1w
Thanks.
TestCodeRepo failures should be fixed CL 767543.dmitshur reviewed +2 on cmd/go/internal/workcmd: include toolchain and godebug in -json1w
I cc'ed you on another CL 724260 that was sent for the same issue earlier, but this seems like a superset of that.
Also noting that the `Directive` definition at https://go.dev/ref/mod#go-work-file-…dmitshur commented on internal/workcmd: add toolchain to json output1w
(1 comment)
Oh, I noticed on the [issue](https://github.com/golang/go/issues/75941#issuecomment-4237842457) that @sean@liao.dev recently sent CL 766480, which already has this comment addressed, and…dmitshur reviewed +2 on internal/workcmd: add toolchain to json output1w
I can review this CL. Thanks for preparing it.
I'll note that in that issue description it's apparent that the 'godebug' is also missing, so this could be changed to 'For #75941' to keep the issue o…dmitshur commented on golang.org/x/build/internal/releasetargets: update list for Go 1.271w
(1 comment)
Oh, I also realized dropping 1.24 release targets will make these tests fail, since they still target that version (as in https://ci.chromium.org/b/8684329214885421377). You could leave …dmitshur opened a change golang.org/x/build/internal/task: fix 'gcloud storage' migration in publishPackageExtension1w
The 'storage' argument here was missed in CL 749081.dmitshur commented on golang.org/x/website/_content/rebuild: mention git dependency1w
Thanks.dmitshur commented on golang.org/x/build/cmd/gorebuild: add a git clone-based fallback to Gerrit's +archive URL1w
Thanks.dmitshur opened a change golang.org/x/website/_content/rebuild: mention git dependency1w
Having git is needed to enable the fallback when fetching directly over
HTTPS runs into insufficient quota and rate limiting.
For golang/go#78804.dmitshur commented on golang.org/x/build/cmd/gorebuild: add a git clone-based fallback to Gerrit's +archive URL1w
(1 comment)
Hmm, it seems this flag is a bit too new e.g. to be included in latest stable Ubuntu, so I'll defer using it. Done in PS 3.dmitshur opened a change golang.org/x/build/cmd/gorebuild: add a git clone-based fallback to Gerrit's +archive URL1w
If the +archive endpoint responds with 429, wait a bit and try again
by using git clone instead. For the purposes of gorebuild it is okay
to wait and make requests slowly: we generally want to run go…dmitshur opened an issue golang.org/x/build/cmd/gorebuild: may fail to produce a useful report due to running into '429 Too Many Requests' unauthenticated rate limiting1w
Gorebuild fetches Go source code from its canonical source, the https://go.googlesource.com/go git repository, and proceeds to check that it can be used to reproduce the associated files posted to ht…dmitshur opened a change golang.org/x/website/_content/doc: add go1.27.md placeholder1w
Add this file now that it's possible to preview the release notes draft
at https://tip.golang.org/doc/go1.27.
For golang/go#78779.dmitshur opened an issue doc: write release notes for Go 1.271w
This is the tracking issue for writing the Go 1.27 Release Notes. The version at tip can be viewed at https://tip.golang.org/doc/next.
When we have a fairly complete release notes draft, enough for …