Activity
Yesterday
dmitshur closed a change runtime/atomic: fix TestAnd64 and TestOr64 on 32-bit architectures23h
Thanks for sending this change. CL 581916 ended up applying the same change as here, so there's nothing more to do here.dmitshur reviewed +2 on testing: use time.Since in highPrecisionTime1d
dmitshur reviewed +2 on testing: use time.Since in highPrecisionTime1d
Thanks.
Please add a mention of the tracking issue, so this change will be easier to find:
```
For #31160.
```dmitshur commented on testing: use QueryPerformanceCounter on Windows1d
(1 comment)
Is it intentional this is implemented as `time.Now().Sub(b.now)` instead of `time.Since(b.now)`? As far as I can tell, that skips over the optimization in https://cs.opensource.google/go…dmitshur reviewed +2 on go.dev/wiki/CodeReview.md: remove RELNOTE=1d
Thanks.dmitshur reviewed +2 on golang.org/x/build/cmd/relnote: improve todo subcommand1d
Thanks.
If you install the latest version of the codereview hooks, and amend:
```
go install golang.org/x/review/git-codereview@latest && \
git codereview hooks && \
git commit --amend --no-edit
``…Last Week
dmitshur closed an issue golang.org/x/build/devapp: support R= annotations in commit messages1d
The Gerrit hashtag `wait-release` has replaced the old R=go1.N syntax (see https://github.com/golang/go/issues/24836#issuecomment-497090827), and so R= isn't used anywhere by now. Issue #24836 tracks…dmitshur commented on golang.org/x/build/cmd/relnote: improve todo subcommand1d
(1 comment)
Probably a good time to update the package comment to refer to its new behavior.dmitshur commented on golang.org/x/build/cmd/relnote: improve todo subcommand1d
As a first pass, looks good. Tests don't build now; I'll look again after you update it for that.
```suggestion
if rn == "yes" || rn == "y" {
```
To match https://cs.opensource.google/go/x/bui…dmitshur closed an issue doc: write release notes in the same CL as the code change1d
This is implemented as of the Go 1.23 cycle by @jba, see issue #64169 and a [golang-dev thread](https://groups.google.com/g/golang-dev/c/Ia9VoYBTRoI).dmitshur commented on golang.org/x/build/cmd/makemac: add first security macOS trybot2d
(1 comment)
Done. Also granted access to makemac's service account.dmitshur commented on golang.org/x/build/cmd/makemac: add first security macOS trybot2d
(1 comment)
Adding a hold to create these secrets.dmitshur opened a change golang.org/x/build/cmd/makemac: add first security macOS trybot2d
For golang/go#65241.dmitshur opened a change golang.org/x/build/main.star: add security-only known issues2d
The next CL adds the darwin-amd64_14 builder type to the security trybot
environment. I'd like it to have a known issue set to begin with, so add
a SECURITY_KNOWN_ISSUE_BUILDER_TYPES dictionary and p…dmitshur reviewed +2 on doc/README.md: discourage RELNOTE in CLs2d
Thanks.dmitshur opened an issue golang.org/x/build: add macOS security trybots to LUCI2d
We have macOS builders in LUCI, but they're not yet available as security trybots. This is the tracking issue for that.dmitshur closed an issue golang.org/x/build/cmd/gitmirror: service problem on 2024-04-232d
dmitshur commented on golang.org/x/build/cmd/gitmirror: service problem on 2024-04-232d
That error doesn't provide detail but suggests an internal problem in git, possibly related to insufficient resources. We hadn't increased gitmirror's resources in a while.
That said, after redepl…dmitshur opened an issue golang.org/x/build/cmd/gitmirror: service problem on 2024-04-232d
Reporting for posterity and in case this comes up again. https://farmer.golang.org/#health was reporting a problem with Git mirroring on 2024-04-23:
. So I agree it seems this needs more investigation. Also see past discussion in https://github.com/glfw/glfw/issues/2133 and htt…dmitshur commented on runtime: add missing runtime error prefix to PanicNilError4d
That meant to say issue #63813. (CL 538496 is this CL.)dmitshur commented on runtime: add missing runtime error prefix to PanicNilError4d
Adding Hold while issue #538496 is in NeedsDecision state.dmitshur commented on os/exec: calling Cmd.Start after setting Cmd.Path manually to absolute path without ".exe" no longer implicitly adds ".exe" in Go 1.224d
@joedian It was a draft CL associated with a review comment. Russ has [go.dev/cl/575255](https://go.dev/cl/575255) which is still open but needs attention.dmitshur commented on go/format: line comment not indented properly after indented statement4d
This is similar to (or the same as) #12822. Also consider this instance:
```Go
switch {
case true:
// comment 1
// comment 2
// comment 3
// comment 4
// comment 5
// comment 6
case fa…dmitshur reviewed +2 on github.com/go-gl/glfw: Remove warning/errors about undefined functions4d
Seems okay to me.
We had this in v3.2 but didn't propagate it forward to v3.3:
https://github.com/go-gl/glfw/blob/a69d953ea14231b2f19998f0b434b6c1e6c6f710/v3.2/glfw/build.go#L25-L26
(It was …dmitshur commented on github.com/sublimehq/sublime_merge: `safe.bareRepository = explicit` makes it impossible to "Edit Commit Message"4d
This still happens with Build 2096 for me.
I also noticed something strange: this happens reproducibly as soon as I change safe.bareRepository to explicit on my primary macOS machine (so I'm unabl…dmitshur commented on cmd/compile: changes to add structs.HostLayout4d
(2 comments)
The 'Minor changes to the library' section is intended for changes to existing packages. Something like a new package being added should go under 'Core library' section, similar to `doc…dmitshur closed an issue golang.org/x/vuln: TestGoModTidy failures5d
I've tried to de-duplicate this into #58766 by updating its pattern to also match this failure.dmitshur commented on golang.org/x/vuln: TestGoModTidy failures5d
This is the same failure mode as in #58766. It's a problem only in the old build infrastructure.dmitshur commented on std/jpeg: ignore garbage bytes before a RST marker5d
(1 comment)
(nit) image/jpeg instead of std/jpeg.dmitshur reviewed +2 on golang.org/x/build/cmd/relui: allow picking which of the private go projects to build from5d
Still looks good, thanks.dmitshur closed an issue golang.org/x/build: gotip-windows-arm64 builders stops working5d
From what I can tell, SWARMING_NEVER_REBOOT has effect for most frequent reasons that would otherwise cause the reboot to happen, but it doesn't catch all. The swarming bot seems to occasionally trig…dmitshur reviewed +2 on golang.org/x/build/main.star: add builders with the c3h88 suffix for highcpu perf tests5d
Thanks.dmitshur reviewed +2 on golang.org/x/build/main.star: introduce and use c2s16 host suffix for performance builders5d
Thanks.
The "c2s16" suffix seems to fall fairly neatly into the CPU selection category (it has to do with performance consistency).dmitshur commented on golang.org/x/build: include all first-class ports in TryBot coverage for x repos against previous Go releases6d
[CL 568197](https://go.dev/cl/568197) added previous Go release coverage for 3 more first-class ports: linux/386, darwin/amd64, and windows/amd64.Earlier
dmitshur starred github.com/emer/gosl/v21w
dmitshur commented on access: Application to join golang/mips maintainer team1w
Also CC @golang/mips (for completeness; almost everyone in it was already CC'ed).dmitshur commented on golang.org/x/build/cmd/watchflakes: start posting1w
Tested this out now; no issues with the fine-grained token permissions so far. I'll keep an eye on it for some time. Submitting.dmitshur commented on all: announce end of support for old macOS releases1w
[CL 563856](https://go.dev/cl/563856) added a note to Go 1.23 release notes, moving to the next.dmitshur commented on golang.org/x/build/cmd/watchflakes: add deployment files1w
Thanks.
Yeah, I gave it fairly minimal resources to start. It mostly makes API calls in a loop, so it might not need much, and we can bump it later as needed. So far it's taking around 7-8 min per i…dmitshur commented on doc: make symbol links simpler for relnotes1w
Thanks.dmitshur commented on golang.org/x/build/cmd/watchflakes: use github.NewClient API1w
We ended up switching to sending only the token in an Authorization header ([docs](https://docs.github.com/en/graphql/guides/forming-calls-with-graphql#communicating-with-graphql)); updated according…dmitshur opened a change golang.org/x/build/cmd/watchflakes: use github.NewClient API1w
Thanks to Russ for adding it.
For golang/go#65353.dmitshur commented on golang.org/x/build/cmd/watchflakes: fetch GitHub token from Secret Manager1w
(3 comments)
We can relax it, but perhaps let's wait before doing it. I added a test so it's easy to spot and to fix if it breaks. I think it'll break now no sooner than net/http changes (e.g., goin…dmitshur reviewed +2 on doc: simplify symbol links for relnotes1w
Thanks.dmitshur merged a change dmitri.shuralyov.com/scratch: hello: change salutations1w
dmitshur reviewed +2 on dmitri.shuralyov.com/scratch: hello: change salutations1w
Someone’s a morning person.
Thanks.dmitshur starred github.com/nishanths/exhaustive1w
dmitshur opened an issue crypto/internal/mlkem768: TestPQCrystalsAccumulated fails on linux/s390x2w
[CL 550215](https://go.dev/cl/550215) added a new package. One of its tests, `TestPQCrystalsAccumulated`, is failing on the linux/s390x port:
```
--- FAIL: TestPQCrystalsAccumulated (0.07s)
…dmitshur commented on access: may-start-trybots and approvers2w
@panjf2000 The Auto-Submit+1 vote is documented at https://go.dev/wiki/GerritAccess#auto-submit. I meant you can ask the person who left the +2 vote to either submit or leave Auto-Submit+1 vote—but…dmitshur starred fyshos.com/fynedesk2w
dmitshur closed an issue access: may-start-trybots and approvers2w
Hi @panjf2000, thanks for contributing to the Go project and asking for approvers access again.
We considered this request. As documented at https://go.dev/wiki/GerritAccess, Gerrit "approvers" ac…dmitshur closed an issue time: clean up duplicate definitions `time.runtimeTimer` and `runtime.timer` that need to be kept in sync2w
This is fixed via [CL 568339](https://go.dev/cl/568339), which removed the duplication.dmitshur closed an issue access: go-assignees and go-editors2w
Hi @qiulaidongfeng, thanks for your interest and contributions to Go.
We considered this request. Similarly to https://github.com/golang/go/issues/66611#issuecomment-2052070282, when considering g…dmitshur closed an issue access: +2 approval in CLs2w
Hi @mauri870, thanks for contributing to the Go project and asking for this access.
We considered this request and felt your contribution history demonstrates enough signals for us to be close to …dmitshur closed an issue access: approvers2w
Hi @qiulaidongfeng, thanks for your interest and contributions to Go.
We considered this request (previously considered in #64255). As documented at https://go.dev/wiki/GerritAccess, Gerrit "appro…dmitshur closed an issue access: Requesting Try-Bots Access and Self-Assign Github Issues2w
Hi @raghvenders, thanks for your contributions to Go so far.
We discussed this request. We generally look for contributors to have a more established track record of contributions before consideri…dmitshur closed an issue golang.org/x/build/cmd/gomote: install builder dependencies by default2w
This is resolved as of [CL 569875](https://go.dev/cl/569875). (It didn't get closed automatically because the Fixes line didn't specify the `golang/go` repo quite right.)dmitshur starred github.com/aclements/go-obj2w
dmitshur commented on golang.org/x/net/http/httpguts: validation of header names, methods, and cookie names could be about twice as fast2w
@jub0bs Yes, your understanding is right and generally it's good to allow some time for discussion on the issue first, though it's still okay to send a change sooner. There are some more details abou…dmitshur starred github.com/cloudpilot-emu/cloudpilot-emu2w
dmitshur starred github.com/migueletto/PumpkinOS2w
dmitshur commented on golang.org/x/build: add LUCI wasip1-wasm32-wasmtime builder3w
Thanks. Confirming a few details.
`GOARCH` should be `wasm32` instead of `wasm`, is that right?
Related to the above—will `GOROOT/misc/wasm` be used for both `wasm` and `wasm32` exec helpers? Tha…dmitshur starred github.com/FiloSottile/FiloSottile3w
dmitshur closed an issue access: may-start-trybots3w
Hi abner chenc,
Thanks for your contributions and helping maintain the linux/loong64 port. We've added you to the `project-golang-may-start-trybots` group, so you should be able to trigger [trybot…dmitshur commented on net/http: TestRequestLimit/h2 becomes significantly more expensive and slower after x/net@v0.23.0 [1.22 backport]3w
> should backport the fix as well
To clarify, do you think the fix to be backported should include changes in [CL 576756](https://go.dev/cl/576756) and [CL 576895](https://go.dev/cl/576895)?dmitshur commented on go/types: issue with using const with implicit unint643w
CC @griesemer, @findleyr.dmitshur commented on golang.org/x/website: release notes for Go 1.22 mentions inexistent math/rand/v2.Uint()3w
Thanks for reporting this.
CC @rsc.dmitshur commented on cmd/compile: unnecessary runtime.growslice with provable capacity3w
CC @golang/compiler.dmitshur commented on cmd/compile: unnecessary bounds check with slice offset3w
CC @golang/compiler.dmitshur commented on go/types: Identical function doesn't work for SSA loaded type3w
CC @griesemer, @findleyr.dmitshur commented on golang.org/x/build: add LUCI linux-ppc64_power8 builder3w
Here's the certificate: [linux-ppc64-power8-1712341779.cert.txt](https://github.com/golang/go/files/14889449/linux-ppc64-power8-1712341779.cert.txt).
I'll also work on another CL needed for that h…dmitshur commented on golang.org/x/build: add LUCI linux-ppc64-power10 builder.3w
@pmur Should this issue also be closed, or is there more to do for it?dmitshur commented on golang.org/x/build: add LUCI linux-ppc64le-power10 builder3w
Thank you very much @pmur. I've mailed you [go.dev/cl/576917](https://go.dev/cl/576917) to remove its known issue.dmitshur opened an issue cmd/coordinator/internal/lucipoll: occasionally gets stuck (for upwards of 24 hrs)3w
The `pollLoop` method that runs in a loop usually takes a minute, but sometimes seems to get stuck and eventually completes in just over 24 hours without any error. This is a tracking issue to sort t…dmitshur opened an issue golang.org/x/tools/gopls/internal/test/marker: 3w
The definition/embed.txt, definition/misc.txt, and hover/sizeoffset.txt cases of `Test` are failing on linux/arm (32-bit) since they were introduced in [CL 573076](https://go.dev/cl/573076).
For e…dmitshur commented on net/http: data race in multiple tests3w
The two instances above were on release-branch.go1.22.dmitshur commented on net: ListenPacket cannot receive packets from non-localhost3w
CC @golang/openbsd, @ianlancetaylor, @neild.dmitshur commented on log/syslog: generated syslog format is not RFC-compliant3w
Thanks for reporting.
As a side note, the [package comment](https://pkg.go.dev/log/syslog#pkg-overview) states the `log/syslog` package is frozen and that some external packages provide more funct…dmitshur commented on golang.org/x/build: add LUCI linux-ppc64le-power10 builder3w
Here's the certificate for the bot hostname `linux-ppc64le-power10`:
[linux-ppc64le-power10-1712180608.cert.txt](https://github.com/golang/go/files/14858800/linux-ppc64le-power10-1712180608.cert.t…dmitshur commented on golang.org/x/build: add LUCI linux-ppc64le_power9 builder3w
Thanks. Here's the certificate: [linux-ppc64le-power9-1712180354.cert.txt](https://github.com/golang/go/files/14858793/linux-ppc64le-power9-1712180354.cert.txt).dmitshur opened an issue net/http: TestRequestLimit/h2 becomes significantly more expensive and slower after x/net@v0.23.03w
I've initially observed this on various builders, most often linux-386 ones, but it seems to reproduce on `darwin/arm64` too.
At tip without x/net@v0.23.0 (e.g., commit 61a3ee54411086d4bf996b65dd3…dmitshur commented on net/http: close connections when receiving too many headers (CVE-2023-45288)3w
@geofffranks There will be a new golang.org/x/net tag with the fix in [CL 576155](https://go.dev/cl/576155) included, soon.dmitshur closed an issue security: fix CVE-2023-45288 [1.21 backport]3w
Closed by merging commit ae5913347d15cf7d1f218916c22717e5739a9ea3 ([CL 576075](https://go.dev/cl/576075)) to `release-branch.go1.21`.dmitshur commented on os/exec: Can't use os.Command (with ExtraFiles) that uses /proc/self/fd/<exe fd> as the executable3w
CC @bradfitz, @ianlancetaylor.dmitshur commented on golang.org/x/build: add LUCI linux-ppc64le-power10 builder3w
Thanks for the ping @pmur. I have time to help move this forward this week.
I've mailed [CL 575856](https://go.dev/cl/575856) as the next step here, which should make it possible for you to start …dmitshur commented on cmd/compile: ICE with `assertion failed` when using `GODEBUG=gotypesalias=1`3w
This seems to also happen at tip (commit 94dba61276f26c01b7d3d1a4abb195bdcc596bfd) if attempting to make.bash with `GODEBUG=gotypesalias=1`.
<details>
```
src $ git pull --ff-only
Already up…dmitshur commented on golang.org/x/build/internal/installer/windowsmsi: Windows installer always adds %USERPROFILE%\go\bin to PATH even if user overrode GOPATH to point to another directory3w
The Windows installer currently always takes responsibility for setting the `GOPATH` environment variable (to `%USERPROFILE%\go`) and adding `%GOPATH%/bin` to `PATH`. If a user wishes to use a differ…dmitshur commented on golang.org/x/build/internal/installer/windowsmsi: Windows installer always adds %USERPROFILE%\go\bin to PATH even if user overrode GOPATH to point to another directory3w
CC @golang/release.dmitshur commented on net/http: HTTP Client slows down over time until stop3w
CC @neild.dmitshur closed an issue reflect/type.go: reflect.StructTag Get does not return 1st tag of dynamic structure created with reflect.StructOf3w
Thanks for reporting. https://pkg.go.dev/reflect#StructTag documentation says:
> By convention, tag strings are a concatenation of optionally space-separated key:"value" pairs. Each key is a non-e…dmitshur reopened an issue cmd/compile: merge temporaries and named variables by GC shape4w
Reopening since the fix in [CL 553055](https://go.dev/cl/553055) was rolled back in [CL 575236](https://go.dev/cl/575236).dmitshur reopened an issue cmd/compile: package-level variable initialization order looks not correct4w
Reopening since the fix in [CL 575175](https://go.dev/cl/575175) was rolled back in [CL 575316](https://go.dev/cl/575316).dmitshur closed an issue cmd/go: spurious "v1.x.y is not a tag" error when a tag's commit was previously download without the tag [1.20 backport]1mo
Go 1.20 has exited its support window per https://go.dev/doc/devel/release#policy.dmitshur closed an issue net/http: don't reuse connections that are experiencing errors [1.20 backport]1mo
The upstream issue isn't resolved yet, and Go 1.20 has exited its support window per https://go.dev/doc/devel/release#policy.dmitshur opened an issue os/exec: calling Cmd.Start after setting Cmd.Path manually to absolute path without ".exe" no longer implicitly adds ".exe" in Go 1.221mo
https://go.dev/doc/go1.22#os/exec includes:
> On Windows, `Command` and `Cmd.Start` no longer call `LookPath` if the path to the executable is already absolute and has an executable file extension…dmitshur opened an issue golang.org/x/build: Windows builders failing with "rdb: The system cannot find the file specified."1mo
We have a better understanding of the problem by now, and a mitigation is in the process of being rolled out.dmitshur closed an issue security: fix CVE-2023-452881mo