Activity
Today
dmitshur commented on : CVEs still showing up in latest version of Go 1.22.225m
Note that Go 1.22.2 is not the current latest version. That version is Go 1.22.3, released 10 days: https://groups.google.com/g/golang-announce/c/wkkO4P9stm0. Go 1.22.3 includes fixes for CVE-2024-24…dmitshur commented on golang.org/x/build: add LUCI freebsd-riscv64 builder1h
From looking at https://ci.chromium.org/ui/p/golang/g/port-freebsd-riscv64/builders, the builder seems to be producing good testing signal. @mengzhuo Would you like to remove its [known issue](https:…dmitshur closed an issue access: Application to join golang/mips maintainer team2h
dmitshur commented on golang.org/x/build/cmd/golangbuild: reduce network bandwidth requirements for builders accessing through LUCI2h
Thanks for reporting. CC @golang/release, @mknyszek.
Are you able to narrow down more specifics of where the additional bandwidth utilization comes from? Is it uplink or downlink?
The testing s…dmitshur commented on time: time.Parse accepts out of range value for offsets, and can't always be roundtripped2h
CC @rsc, @ianlancetaylor.dmitshur commented on unsafe: document a difference in behavior between unsafe.Sizeof(x) and reflect.TypeOf(x).Size()2h
CC @griesemer.dmitshur commented on golang.org/x/review/git-codereview: set git config commit.gpgsign to false2h
Thanks for reporting this.
The git configuration surface is quite large, and this appears to be a case where one of the configuration values conflicts with the expectations of git-codereview, caus…dmitshur commented on cmd/dist: Go tip leaving pkg/mod/golang.org/x/telemetry read-only files behind2h
Thanks for reporting. Some notes from when I briefly looked in case they're helpful.
cmd/dist sets GOPATH env var (which in turn sets the module cache location) to $GOROOT/pkg/obj/gopath with the …dmitshur commented on crypto/tls: move defaults into defaults.go and defaults_boring.go3h
Let's use Commit-Queue+1 to trigger trybots in this CL. See https://go.dev/wiki/LUCI#trybots.dmitshur reviewed +2 on golang.org/x/tools/...: fix function names in comment3h
Thanks.
@imcusg@gmail.com Please see https://go.dev/wiki/LUCI#trybots for current information on how to trigger trybots. In most cases you should use Commit-Queue+1 vote, not the legacy Run-TryBot+1…dmitshur commented on gerrit: Unable to select specific builder to execute tryjobs3h
Thanks for reporting. CC @golang/release.
First, can you try and see if you're able to start slowbots by including a `Cq-Include-Trybots: luci.golang.try:gotip-linux-loong64` git footer in the com…dmitshur commented on access: merge my Gerrit accounts4h
Sorry for the trouble, I didn't realize it wouldn't have effect on your newer CLs.
Since it's only two of them and both haven't begun code review yet (and one of them needs a merge conflict resolu…dmitshur commented on golang.org/x/build: add LUCI illumos-amd64 builder11h
Thanks for getting to this point.
The bot is visible at https://chromium-swarm.appspot.com/bot?id=illumos-amd64-jclulow. As you noted, it's identifying itself as having a cipd_platform value of "s…Yesterday
dmitshur commented on golang.org/x/build: LUCI support for plan915h
Thank you for working on this and your patience. I'd like to try to help move this forward. Apologies for not commenting here sooner.
We don't quite have an ideal existing builder that we can poin…dmitshur commented on golang.org/x/build/env: host-linux-amd64-androidemu (Container Image: android-amd64-emu:bff27c0c9263)15h
This particular issue tracks the Android emulator-based builder. I've had a chance to prototype it earlier, and will update it as further progress is made.
@changkun, at this time this issue doesn…dmitshur commented on net/http: return correct error when reading from a canceled request body16h
(1 comment)
Consider adding a `Cq-Include-Trybots: luci.golang.try:gotip-linux-amd64-longtest` footer (detailed instructions [here](https://go.dev/wiki/SlowBots#using-slowbots)) so trybot signal wo…dmitshur commented on net/http: return correct error when reading from a canceled request body16h
(1 comment)
I retitled the issue to be about net/http.dmitshur reviewed +1 on net/http: return correct error when reading from a canceled request body16h
Looks good to me, but I'll leave it to Jonathan to look, so +1. Left a few minor comments that I spotted. Thanks.
(nit) https://go.dev/wiki/Spelling suggests using a different spelling for this part…dmitshur commented on golang.org/x/build/buildlet: don't rely on io.Copy error to be context.DeadlineExceeded16h
Hold because CL 586315 will restore the net/http behavior, and this shouldn't be needed.dmitshur commented on golang.org/x/sys/windows: uintptr(unsafe.Pointer) calls outside of syscalls17h
CC @golang/runtime, @alexbrainman.dmitshur commented on cmd/go: go 1.22.3 go test -cover failed result17h
CC @thanm.
dmitshur commented on golang.org/x/net/ipv4: calling SetTOS and SetTTL doesn't work as expected on MacOS17h
CC @golang/darwin, @ianlancetaylor.dmitshur commented on golang.org/x/playground: corrupted output on long slice printing17h
CC @golang/tools-team.dmitshur commented on golang.org/x/build/buildlet: don't rely on io.Copy error to be context.DeadlineExceeded18h
(1 comment)
This is dead code now; removed.dmitshur closed an issue golang.org/x/build/buildlet: add context support18h
We can call this done. At least no use in tracking this via an open issue anymore.dmitshur commented on golang.org/x/build/buildlet: TestExecTimeoutError fails consistently after CL 546676 ("net/http: simplify HTTP/1 request cancelation")18h
I sent [CL 586077](https://go.dev/cl/586077) which fixes the immediate problem on the x/build side. It'd be good for Damien to check if this is working as expected.dmitshur opened a change golang.org/x/build/buildlet: don't rely on io.Copy error to be context.DeadlineExceeded18h
Previously it worked for Exec to detect that the context deadline
exceeded by inspecting the io.Copy error. As of CL 67439, that no
longer works, so add an explicit check if the context exceeded if
i…dmitshur commented on net/http: simplify HTTP/1 request cancelation18h
FYI an old test in x/build has started to fail with this change. I filed #67439 and will see if I can spot a problem on the x/build side.dmitshur opened an issue golang.org/x/build/buildlet: TestExecTimeoutError fails consistently after CL 546676 ("net/http: simplify HTTP/1 request cancelation")18h
As visible on the [x/build by Go commit](https://ci.chromium.org/p/golang/g/x-build-gotip-by-go/console) view, `TestExecTimeoutError` has started to fail consistently at Go tip starting with [CL 5466…dmitshur reviewed +2 on crypto/x509: [release-branch.go1.22] remove TestPlatformVerifierLegacy tests20h
Thanks.dmitshur reviewed +2 on crypto/x509: [release-branch.go1.21] remove TestPlatformVerifierLegacy tests20h
Thanks.dmitshur commented on golang.org/x/build: add LUCI illumos-amd64 builder20h
Thank you for starting to work on this builder. Here's the certificate:
[illumos-amd64-jclulow-1715889249.cert.txt](https://github.com/golang/go/files/15340192/illumos-amd64-jclulow-1715889249.cert…dmitshur commented on golang.org/x/build/cmd/gopherbot: add task to label access issues22h
Thanks.dmitshur commented on crypto/x509: cleanup signature generation22h
Note that the parent commit is almost 1k commits old, which is much for a bigger change like this, causing trybot results to be less very representative. I'll rebase and restart trybots, just in case…dmitshur closed an issue access: merge my Gerrit accounts22h
In this case, we can merge your two accounts into one. That is done now.
Specifically, both of your emails are now associated with your older Gerrit account (ID 61713, created Mar 10, 2023), https…dmitshur commented on cmd/go: unexport loadImport23h
(1 comment)
Since GOPATH-based "go get" is no more, should this be removed?dmitshur commented on access: may-start-trybots23h
To confirm, you are able to leave a Commit-Queue+1 vote using your account, is that right? I see [here](https://go-review.googlesource.com/c/go/+/544435/1#message-07e9ce8108aeb839e2e9be291985b79117b0…dmitshur reviewed +2 on doc/next/64910.md: reword release note23h
Thanks.dmitshur reviewed +2 on doc/README.md: give example of package-less symbol link23h
Thanks.dmitshur reviewed +2 on doc: make symbol links simpler for relnotes1d
Thanks.
Great. That is simpler.dmitshur commented on doc: make symbol links simpler for relnotes1d
Thanks, this looks good, just wanted to ask one question first.
Do you know if this would link correctly if this is written as [encoding.BinaryMarshaler]? (You can check by looking at `relnote gener…dmitshur commented on access: may-start-trybots1d
The gerrit account `chenguoqi@loongson.cn` is still present in the `project-golang-may-start-trybots` group, so you should be able to. Are you logged in with that account and following the directions…dmitshur commented on access: Application to join golang/mips maintainer team1d
Thank you for contributing, and sorry this access request took us a while to process. You should have a pending invite to join the golang/mips team.dmitshur opened a change golang.org/x/build/cmd/gopherbot: add task to label access issues1d
This label will be used to automatically add new access issues to the
incoming queue.
Tested with go run . -dry-run -only-run='label access issues', result
is as expected.dmitshur commented on crypto/x509: cleanup signature generation1d
(1 comment)
FYI you can now write this in a simpler way as [CreateCertificateRequest]. It will be auto-linked as an identifier in this same package, like in godoc.
See [doc/README.md](https://cs.op…dmitshur commented on net/http: TestServerReadAfterWriteHeader100Continue/h1 failures1d
CC @neild.dmitshur opened a change golang.org/x/build/internal/relui/sign: re-enable non-existent signing request test1d
Take measures to wait for fakeSigningServerClient to complete before
returning from a test (so it can't call t.Logf after the test exited).
Also avoid calling t.FailNow inside goroutines other than t…dmitshur closed an issue golang.org/x/build/internal/relui/sign: TestUpdateSigningStatus failures1d
Fixed by [CL 504495](https://go.dev/cl/504495).dmitshur opened a change golang.org/x/build/internal/task: delete constant timeouts to deflake tests1d
The motivation is the same as described in the commit message
of https://go.dev/cl/504495.
For golang/go#64110.
Fixes golang/go#66282.dmitshur closed an issue build: build failure on x_vulndb-gotip-darwin-amd64_131d
Duplicate of #67404dmitshur closed an issue build: build failure on x_tools-gotip-solaris-amd641d
Duplicate of #67404dmitshur closed an issue build: build failure on x_tools-gotip-darwin-amd64-longtest1d
Duplicate of #67404dmitshur closed an issue build: build failure on x_vulndb-gotip-darwin-amd64_121d
Duplicate of #67404This Week
dmitshur closed an issue build: build failure on x_vulndb-gotip-darwin-amd64_111d
Fixed by [CL 585818](https://go.dev/cl/585818).dmitshur commented on net/http: avoid panic when writing 100-continue after handler done1d
(1 comment)
This test appears to be flaky. See #67382.dmitshur commented on golang.org/x/build: add LUCI freebsd-riscv64 builder1d
There's now a fully passing build at https://ci.chromium.org/b/8747907145247685825!dmitshur closed an issue golang.org/x/build/cmd/watchflakes: opens duplicate issues1d
We've made two changes to help with this:
- there's now an auto-archive workflow in the Test Flakes project
- it archives closed issues after 4 weeks of inactivity; we can adjust the exact ti…dmitshur closed an issue net/http: flaky test timeout1d
This issue isn't in the Test Flakes project, so it's not receiving further updates from watchflakes. If there are matches, they're being reported elsewhere. Closing this as there's nothing more to do…dmitshur commented on github.com/golang/protobuf: Set go.mod Go language version to oldest supported Go version (1.21 currently)1d
@jhump Thanks for your comment. I want to address the second paragraph in https://github.com/golang/protobuf/issues/1613#issuecomment-2113378587: the golang.org/x/... modules are only tested with the…dmitshur commented on cmd/cgo: a type alias of unsafe.Pointer is not treated as unsafe.Pointer1d
@ianlancetaylor Assigned it to you for now, please feel free to update.dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383dmitshur closed an issue runtime:cpu4: TestUsingVDSO failures1d
Duplicate of #67383.dmitshur commented on golang.org/x/build: add LUCI openbsd-ppc64 builder1d
Thanks for working on this.
> I'm not illuminated by looking at the Ended Builds list either.
I failed to realizer this sooner, but our configuration intends to make it possible for you to see …dmitshur commented on crypto/x509: TestPlatformVerifier failures on Windows due to broken connections [1.22 backport]1d
This is now approved for backporting. @rolandshoemaker Is it convenient for you to mail the backport CLs? Thanks.dmitshur commented on cmd/go: [release-branch.go1.21] set global .gitconfig location more robustly in tests1d
The TestServerGracefulClose/h1 failure looks like a pre-existing flake, not newly caused by changes in this CL. All other tests passed. Adding bypass.dmitshur commented on net/http: nil pointer dereference panic for httputil.ReverseProxy users [1.21 backport]1d
See https://github.com/golang/go/issues/67375#issuecomment-2113061816.dmitshur commented on net/http: nil pointer dereference panic for httputil.ReverseProxy users [1.22 backport]1d
We've discussed this candidate and are currently leaning towards rejection: this is an issue that has existed since at least Go 1.18, with difficulty in reliably reproducing it and not many users run…dmitshur commented on golang.org/x/build/cmd/watchflakes: opens duplicate issues1d
I've stopped it from running as a short-term workaround. Removing Soon label.
We've applied to https://github.blog/changelog/2024-02-12-github-issues-projects-projects-without-limits-private-beta/…dmitshur commented on golang.org/x/build/cmd/watchflakes: opens duplicate issues1d
Logs show it's running into a limit of items that can be added to a GitHub project:
```
# new issue: runtime:cpu4: TestUsingVDSO failures
2024-05-15 13:52 gotip-linux-ppc64_power8 go@6ccd8e4c ru…dmitshur opened an issue golang.org/x/build/cmd/watchflakes: opens duplicate issues1d
watchflakes is opening many duplicate issues recently, like "runtime:cpu4: TestUsingVDSO failures" is go.dev/issue/67392 and go.dev/issue/67391 and others.
I suspect the problem is that they're no…dmitshur reviewed +2 on doc/README.md: mention adding the proposal issue1d
Thanks. Sorry about the delay.
It certainly makes sense to use the relative '/issue/nnn' path in actual links, since they'll work as intended when they're rendered on each website.
But that doesn't…dmitshur commented on golang.org/x/build/cmd/relnote: missed API changes2d
@jba Checking on from a weekly release meeting. Is there more to do before considering this resolved? Thanks.dmitshur commented on cmd/go: list confuses imports2d
@matloob Are there updates on this? This is currently a blocker for Go 1.23 RC 1. Is this important enough to be considered a release blocker for Go 1.23?dmitshur opened an issue x/build/internal/installer/{darwinpkg,windowsmsi}: report if GOTOOLCHAIN in environment is set to an older version than one being installed2d
When an installer for a certain Go version runs, it can check if the environment is configured such that `go version` would select select an older toolchain than the one being installed, and potentia…dmitshur starred github.com/gen2brain/raylib-go2d
dmitshur commented on golang.org/x/build: add LUCI freebsd-riscv64 builder2d
Looking at https://chromium-swarm.appspot.com/bot?id=freebsd-riscv64-mengzhuo, the bot seems to be running well in many x/ repos. It passes all but 2 tests at tip (e.g., https://ci.chromium.org/b/874…dmitshur commented on golang.org/x/build/cmd/golangbuild: LUCI builders fail with rdb-stream fail to parse FAIL message2d
The problem appears to be in `result_adapter`, which is what's used to parse `go test -json` output and upload it to ResultDB. See:
https://source.chromium.org/chromium/infra/infra/+/main:go/src/i…dmitshur commented on log: Fatal() reports wrong source location when both log.Lshortfile or log.Llongfile and slog with source is enabled2d
CC @robpike.dmitshur commented on os/exec: execution of internal command at Windows gives error2d
This is probably working as intended. If there was a command named `echo` in %PATH%, would it be executed? If so, it might not be a responsibility of os/exec to track which builtins on various platfo…dmitshur commented on os/exec: execution of internal command at Windows gives error2d
CC @golang/windows.dmitshur commented on runtime/cgo: `x_cgo_getstackbound()` broken on SmartOS2d
CC @golang/illumos.dmitshur reviewed +2 on golang.org/x/build/dashboard: update host-linux-loong64-3a5000 owner2d
dmitshur reviewed +2 on golang.org/x/build/dashboard: stop linux-loong64 builders2d
Thanks.dmitshur commented on golang.org/x/website/_content/doc/install.html: eliminate redundant link2d
It failed because the parent commit 01ff1375 is quite old, before CL 516256 fixed the build error on Plan 9. If you rebase to a newer base commit, that should fix it.
That said, other than the build…dmitshur commented on encoding/binary: add Append, Encode and Decode2d
(1 comment)
Should this have 'Fixes #60023.' line here or so?dmitshur commented on cmd/go: turn back on build_plugin_reproducible script test2d
(1 comment)
Yeah, this is tricky. I think we need to find a way to reliably reproduce it, then we can look at what bytes differ and get an idea of where the non-reproducibility comes from.
For now,…dmitshur commented on cmd/go: turn back on build_plugin_reproducible script test2d
(1 comment)
I'll also note that I tried this CL locally on a Mac with the same Xcode version 15.3 (15E204a) and CL 585355 was effective in fixing it, which suggests there's something more than just …dmitshur commented on cmd/go: turn back on build_plugin_reproducible script test2d
(1 comment)
If the result is reproducible, it suggests there's some difference between how the test was run on the gomote and how it was run on the builder. But the builder result is more representa…dmitshur commented on encoding/base32: use correct length for unpadded buffer in Read3d
(1 comment)
Patch set 7 resolved this. Thanks.dmitshur commented on cmd/go: turn back on build_plugin_reproducible script test3d
(1 comment)
Let's try it on the builder:
```
Cq-Include-Trybots: luci.golang.try:gotip-darwin-amd64-longtest
```
(More context on that git footer is at https://go.dev/wiki/SlowBots#reviewer-workfl…dmitshur commented on golang.org/x/build/dashboard: stop linux-loong64 builders3d
You're right that as is this change isn't doing more than removing slowbot aliases. Since the entire coordinator codebase is going away after it's stopped, it doesn't make too much of a difference an…dmitshur reviewed +2 on golang.org/x/build/dashboard: update host-linux-loong64-3a5000 owner.3d
Thanks.
(nit) No need for a trailing period (see https://go.dev/wiki/CommitMessage).dmitshur commented on golang.org/x/build/dashboard: stop linux-loong64 builders3d
(1 comment)
Thanks for making the change. Since this is a reverse builder, it's not necessary to make a change in coordinator to stop it. You can stop it from connecting on your end.
If you do want…