Activity
There was a problem fetching the latest activity.
This Week
dmitshur commented on snappy: tag a final release1d
CC @nigeltao.dmitshur opened an issue github.com/immersive-web/webxr-samples: broken link to WebXR/WebGPU bindings on WebXR Report page2d
Hello, I'm reporting that the 'WebXR/WebGPU bindings' link at https://immersive-web.github.io/webxr-samples/report/ is broken:
https://github.com/immersive-web/webxr-samples/blob/082da254f525a9ec077…dmitshur commented on golang.org/x/tools/gopls/internal/telemetry/cmd/stacks: go build within TestReadPCLineTable fails with "go: unknown GOEXPERIMENT synctest"2d
The recent failures are not the original "unknown GOEXPERIMENT synctest" error, but a different problem, that just happens to be in the same test. Narrowed down the watchflakes query to let it file a…dmitshur commented on cmd/go: packages with non root vendor elements excluded from module3d
Issue #37397 was about some packages whose import path _ends_ with "/vendor" being treated as if they're vendored, and being left out of module zips.
It looks like what you're running into isn't tha…dmitshur commented on github.com/mdempsky/unconvert: Internal error3d
@JetSetIlly Does `go version -m $(which unconvert)` print v0.0.0-20250216222326-4a038b3d31f5 or an older version? The latest tag cache might not have worn off yet, meaning you'd need to do `go instal…dmitshur fixed an issue github.com/mdempsky/unconvert: Internal error3d
dmitshur fixed an issue github.com/mdempsky/unconvert: Internal error3d
dmitshur merged a change github.com/mdempsky/unconvert/...: update go and dependency versions [generated]3d
dmitshur commented on github.com/mdempsky/unconvert/...: update go and dependency versions [generated]3d
Thanks.dmitshur opened a change github.com/mdempsky/unconvert/...: update go and dependency versions [generated]3d
Fixes #66.
[git-generate]
go get go@1.23.0 toolchain@none golang.org/x/tools@v0.30.0 golang.org/x/text@v0.22.0 go mod tidy
sed -i '' 's/21/24/' .github/workflows/test.yml
sed -i '' 's/20/23/' .…dmitshur commented on golang.org/x/net/context: deprecate package or symbols in package3d
There's also a [context fix](https://cs.opensource.google/go/go/+/master:src/cmd/fix/context.go;l=15-21;drc=bf90da97c1aaec78d2f8ad8b74a506d3b6f0ee75) available to aid with the migration, so
[`go fix`…Last Week
dmitshur commented on access: approvers5d
Hi @cpu,
Thank your for contributing to the Go project and all of your work on packages related to Go's security and cryptography. We've discussed this request on the team and agreed to grant you th…dmitshur fixed an issue access: approvers5d
dmitshur commented on github.com/dgraph-io/badger/v4/fix(aix): add aix directory synchronization support5d
In case it's helpful, I'll point out that you can get some signal from compile-only testing via cross-compilation. For example, `GOOS=aix GOARCH=ppc64 go test -c -o=/dev/null` can be run on any platf…dmitshur commented on cmd/go: stamped (pseudo) version on build does not work for module in subdirectory5d
CC @matloob, @samthanawalla.dmitshur commented on cmd/go: go run and go tool are still quite slow (50ms on Mac M1 Max with warm caches)5d
CC @matloob, @samthanawalla.dmitshur commented on all: update standard-library dependencies at the start and end of each development cycle6d
With CL 648735, CL 648916, and CL 649296, everything's updated for the early-in-cycle round. Removing that label to revisit this when the release freeze starts.dmitshur commented on net: Resolver leads to too many DNS queries6d
CC @ianlancetaylor, @neild.dmitshur commented on crypto/tls: wrong error returned in ECHConfig parsing6d
CC @golang/security.dmitshur commented on golang.org/x/build: gotip-linux-amd64-misccompile builder fails with no space left on device1w
Oh, so the $HOME/.cache isn't being cleaned by anything? Yeah, using a cache directory in a per-build work directory that is cleaned sounds good.
Did you mean XDG_CACHE_HOME (s/_DIR/_HOME)? Setting …dmitshur commented on golang.org/x/build: gotip-linux-amd64-misccompile builder fails with no space left on device1w
Thanks for filing an issue to track this.
Oddly, as Michael K. pointed out, it seems this isn't happening in post-submit so far, yet fairly often in pre-submit recently. Compare with:
https://ci.ch…dmitshur commented on golang.org/x/tools/cmd/goimports: add a flag to skip generated files1w
CC @golang/tools-team.dmitshur commented on encoding/asn1: ObjectIdentifier Unmarshal error if SubOID > MaxInt32 on 64bit machine too1w
CC @golang/security.dmitshur commented on runtime: recover calls in loop body of ranging over iterators doesn't take effect1w
CC @golang/runtime.dmitshur commented on runtime: recover added in range-over-func loop body doesn't stop panic propagation / segfaults printing error1w
CC @golang/runtime.dmitshur commented on os: MkdirAll fails when a parent folder contains trailing spaces 1w
@qmuntal If I understand right, when you applied the release-blocker label in Nov 2024 it was meant to be to track reverting CL 618496 in time for Go 1.24, is that right? Note that when using that la…dmitshur commented on build: windows/arm (32-bit ARM) port is broken1w
@mcandre go tool dist list already leaves out broken ports by default. Broken ports can be listed too by passing in the `-broken` flag.
<details><br>
```
$ go version
go version go1.24.0 darwin/arm…dmitshur opened an issue doc: write release notes for Go 1.251w
This is the tracking issue for writing the Go 1.25 Release Notes.
When the Go 1.25 Release Notes are complete, including resolving all known items and those caught by `relnote todo`, the release-blo…dmitshur commented on crypto/tls: missing "configForClient.mutex.RUnlock()" in "Config.ticketKeys(configForClient *Config)" func1w
Thanks for the report. I agree it looks like the code is missing a matching RUnlock call on configForClient.mutex in that case. This code doesn't seem to have changed recently, but it only comes up w…dmitshur fixed an issue cmd/go/internal/vcweb/vcstest: TestScripts/bzr/hello.txt failures1w
dmitshur commented on cmd/go/internal/vcweb/vcstest: TestScripts/bzr/hello.txt failures1w
The builder issue (#71563) is resolved.dmitshur fixed an issue golang.org/x/build/env/linux-arm64-bookworm: restore a working bzr installation1w
dmitshur commented on golang.org/x/build/env/linux-arm64-bookworm: restore a working bzr installation1w
Fixed by [crrev.com/c/6250541](https://crrev.com/c/6250541).dmitshur commented on os: possible to remove or change internal finalizer on a `Root`1w
Closed by merging [CL 648135](https://go.dev/cl/648135) (commit 6d399e9da6090af289aba1f9c4bcc8488387ff9a) to release-branch.go1.24.dmitshur fixed an issue os: possible to remove or change internal finalizer on a `Root`1w
dmitshur reopened an issue os: possible to remove or change internal finalizer on a `Root`1w
Since this release-blocker is in the Go 1.24 milestone, re-opening to track cherry-picking to release-branch.go1.24.Earlier
dmitshur commented on golang.org/x/build: catch problems in x/ repos gated by module go directive version sooner1w
As far as I can tell everything's working as expected. I filed #71613 and #71612 for the current findings.
Is there anything more to do here @mknyszek or ready to call this done?dmitshur opened an issue golang.org/x/build: new linux-arm64-msan-clang15 builder isn't passing at tip yet1w
[CL 643918](https://go.dev/cl/643918) added ASAN and MSAN builders for linux-arm64, related to issue #70054. The gotip-linux-arm64-asan-clang15 builder is passing, but the gotip-linux-arm64-msan-cla…dmitshur opened an issue golang.org/x/tools/cmd/signature-fuzzer: test failures when module's language version is 1.241w
When x/tools' go.mod is updated to Go language 1.24 or newer (which will happen this August based on https://go.dev/s/release#timeline and https://go.dev/design/69095-x-repo-continuous-go#why-1_n_1_0…dmitshur opened an issue golang.org/x/crypto: 4 tests will run into https://pkg.go.dev/crypto/rsa@go1.24rc3#hdr-Minimum_key_size1w
Go 1.24 makes a change to crypto/rsa.GenerateKey that makes it return an error if a key of less than 1024 bits is requested, also adding a GODEBUG rsa1024min for it. See https://go.dev/doc/go1.24#cry…dmitshur commented on cmd/go/internal/vcweb/vcstest: TestScripts/bzr/hello.txt failures1w
This remains a problem only on the release branches now. We'll fix that either by backporting a test-only fix, or more likely fixing the builder soon (tracked in #71563).dmitshur commented on weak: weak.Pointer[T] can be misused with type conversions1w
Closed by merging [CL 647435](https://go.dev/cl/647435) (commit b7b4c60585f7b341762b70a799e854e991e770ac) to release-branch.go1.24.dmitshur fixed an issue weak: weak.Pointer[T] can be misused with type conversions1w
dmitshur reopened an issue weak: weak.Pointer[T] can be misused with type conversions1w
Since this is in the Go 1.24 milestone, re-opening to track cherry-picking to release-branch.go1.24.dmitshur commented on weak: weak.Pointer[T] can be misused with type conversions1w
CC @cherrymui.dmitshur commented on github.com/go-gl/glfw/v3.3/glfw: Annotate functions with #cgo noescape/nocallback where possible2w
Yeah, it’s interesting how much that can help here. We’ll need some measurements to evaluate and decide.
Somewhat separately, it’d be good if we can help land v3.4, and target such enhancement…dmitshur commented on make.bash: build failure due to "undefined: sockaddrLen"2w
Fixed in [CL 646975](https://go.dev/cl/646975).dmitshur fixed an issue make.bash: build failure due to "undefined: sockaddrLen"2w
dmitshur opened an issue golang.org/x/build/cmd/relui: add support for mentioning security fixes in Release Candidate release kind2w
Right now relui supports sending release announcements to the golang-announce mailing list that mention security fixes only for minor Go releases:
https://cs.opensource.google/go/x/build/+/master:in…dmitshur opened an issue golang.org/x/build/env/linux-arm64-bookworm: restore a working bzr installation2w
The linux-arm64-bookworm image installs bzr via apt get, and that installation intends to work with the system python in /usr/bin. However, golangbuild ends up having a different version of the `pyth…dmitshur commented on cmd/go/internal/vcweb/vcstest: TestScripts/bzr/hello.txt failures2w
Removing Soon label since [CL 646635](https://go.dev/cl/646635) added a test skip to allow for more time to investigate this.dmitshur commented on cmd/go: TestScript/cover_coverprofile_nocoverpkg failures due to unrecognized GOEXPERIMENT "coverageredesign" [consistent failure]2w
Fixed by [CL 646355](https://go.dev/cl/646355). (Thanks @ianlancetaylor.)dmitshur opened an issue internal/trace: TestTraceCgoCallback/Default fails on new freebsd-amd64-race LUCI builder with FreeBSD 14.12w
Running this test on a new LUCI builder configuration that sets the `-race` flag on FreeBSD 14.1 fails with:
```sh
# GOOS=freebsd GOARCH=amd64 go test -race -run=^TestTraceCgoCallback$ internal/trac…dmitshur commented on cmd/go: TestScript/version_buildvcs_bzr failures2w
Merged into #71504.dmitshur fixed an issue cmd/go: TestScript/version_buildvcs_bzr failures2w
dmitshur starred github.com/prattmic/benchfmt-convert2w
dmitshur commented on net/http: TestTransportIdleConnTimeout failures3w
Looks like [CL 637897](https://go.dev/cl/637897) indeed fixed this. Closing, watchflakes will follow up if it still happens.dmitshur fixed an issue net/http: TestTransportIdleConnTimeout failures3w
dmitshur commented on golang.org/x/build/cmd/coordinator: older dashboard pages omit LUCI builders in detailed views (page 2, x/ repo commits) but don't make it apparent it's WAI3w
Thanks for reporting this.
The missing builders on those pages are all the new LUCI builders. In #65913 we provided limited support for displaying LUCI build results on build.golang.org, but that su…dmitshur closed an issue golang.org/x/build: Tricium should not enforce Go style rules on testdata/*.go3w
The implementation of the check we were using previously is fully gone as of [CL 643916](https://go.dev/cl/643916). I'll close this since the original problem is no longer happening and so the issue …dmitshur commented on golang.org/x/build: run platform/arch specific builders when platform/arch constrained files are changed4w
To connect some dots: in #66009, @mknyszek has a paragraph that mentions a "proxy" builder, which is similar (if not identical) to the "spawning an additional set of builders" description here.
At a…dmitshur commented on go/types: dot imports might cause errors where `(types.Error).Pos` is not part of the `(types.Error).Fset`4w
@adonovan You have a fix ready, do you intend to land it for Go 1.24 or 1.25? (Asking because the CL doesn't have wait-release hashtag, and there's no milestone set on this issue.) It's quite late in…dmitshur commented on net/http: do not enable HTTP/2 extended connect by default4w
@cdujeu For the golang.org/x/net module specifically, [CL 641475](https://go.dev/cl/641475) has already been submitted, so you're able to upgrade to a new enough version of that module that includes …dmitshur commented on golang.org/x/build: add LUCI linux-s390x builder1mo
@srinivas-pokala Thanks for working on this. I looked at the logs on our end for error details.
I'm seeing 403s that are failing because the bot ID being reported is "go-s390x01" instead of the exp…dmitshur commented on cmd/dist: check that builds are reproducible1mo
The reproducibility of builds is verified during the release process, and the https://go.dev/rebuild reports a daily reproducibility report. There have not been problems reported or uncovered for a w…dmitshur fixed an issue cmd/dist: check that builds are reproducible1mo
dmitshur opened an issue golang.org/x/build/cmd/relui: add support for making release announcement posts on Bluesky1mo
During Go releases, relui takes on posting on 2 social sites using their APIs (#47403 and #65046). By now we have an account on one more social site, Bluesky, where we've also started to post links t…dmitshur commented on golang.org/x/build: catch problems in x/ repos gated by module go directive version sooner1mo
Looks real; it reproduces consistently with go1.24rc2 darwin/arm64. Glad to see it paying off already!
<details>
```
tools $ go test -count=1 ./cmd/signature-fuzzer/internal/fuzz-generator
ok go…dmitshur fixed an issue net/http: sensitive headers incorrectly sent after cross-domain redirect [CVE-2024-45336] [1.24 backport]1mo
dmitshur fixed an issue crypto/x509: usage of IPv6 zone IDs can bypass URI name constraints [CVE-2024-45341] [1.24 backport]1mo
dmitshur closed an issue net: MPTCP support is not enabled when apps use `ListenTCP()` or `DialTCP()` [1.22 backport]1mo
dmitshur closed an issue net: MPTCP support is not enabled when apps use `ListenTCP()` or `DialTCP()` [1.23 backport]1mo
dmitshur commented on golang.org/x/build/cmd/golangbuild: update result_adapter to handle Go 1.24's go test -json output1mo
[crrev.com/c/6174231](https://crrev.com/c/6174231) updates result_adapter with better support for Go 1.24 structured build events.
After that's in, we can stop setting GODEBUG=gotestjsonbuildtext=1 …dmitshur commented on doc: write release notes for Go 1.241mo
@andig The `go help get` text describes the `go get` command and provides help on using it, but it doesn't attempt to provide complete coverage for all details. For that, it refers the reader to refe…dmitshur commented on golang.org/x/build: catch problems in x/ repos gated by module go directive version sooner1mo
Yeah, determining it automatically without configuration, like for perf builders, sounds good. (If we later on realize we want more flexibility for some reason, we can add that flexibility afterwards…dmitshur commented on archive/zip: AddFS omits trailing slash in directory names1mo
CC @songgao, @ianlancetaylor.dmitshur commented on golang.org/x/pkgsite/devtools/cmd/seeddb: -keep_going through version list errors1mo
Note that the module proxy protocol specification (https://go.dev/ref/mod#goproxy-protocol) is clear about expected HTTP status codes:
> Successful HTTP responses must have the status code 200 (OK…dmitshur commented on golang.org/x/tools/go/packages: Load does not respect Go version consistently1mo
@StevenACoffman What you’re describing may not necessarily be the same issue. This issue seems to be about inconsistency in go version used when “go1.N.M” binaries are used to run vs to build G…dmitshur opened an issue golang.org/x/build/cmd/watchflakes: fails to parse the "!~" operator1mo
Both of these scripts this should parse successfully:
```
default <- pkg == "cmd/go" && output ~ `some regexp`
default <- pkg == "cmd/go" && output !~ `some regexp`
```
Yet the latter doesn'…dmitshur commented on cmd/go: TestScript failures1mo
@thepudds Thanks for looking at this too, and suggesting that workaround. I'll use it for now.
I should've posted a comment here sooner, but I think I followed a similar path and ended up with [CL…dmitshur commented on : Vulnerability GHSA-3xgq-45jj-v275/CVE-2024-21538 found in cross-spawn-7.0.31mo
I see it's a dependency in some of the golang.org/x repos:
https://cs.opensource.google/search?q=%22cross-spawn%22%20%227.0.3%22&ss=godmitshur commented on : Vulnerability GHSA-3xgq-45jj-v275/CVE-2024-21538 found in cross-spawn-7.0.31mo
This is the issue tracker for the Go project. What is "cross-spawn" in your report? If it's referring to https://www.npmjs.com/package/cross-spawn, the issue tracker for that seems to be https://gith…dmitshur commented on golang.org/x/build: add LUCI openbsd-riscv64 builder1mo
Okay, let's update the LUCI bot's coverage to match the previous one, then. I'll work on a CL for that.dmitshur commented on golang.org/x/build: update OpenBSD builders to 7.61mo
For the 386/amd64 builders that we run as VMs in GCE, this will first need to go through a process similar to what was recently done in #61095 (e.g., see [CL 616836](https://go.dev/cl/616836) which a…dmitshur commented on cmd/compile: gotypesalias=1 does not seem to get used correctly for deps on Go 1.231mo
The Go 1.23 backport is tracked in another issue, Go 1.24 is already fixed, and I sent [CL 640017](https://go.dev/cl/640017) to remove the 1.23 skip in the x/tools test. @timothy-king Is there anythi…dmitshur commented on crypto/tls: TestVerifyConnection/TLSv12 failures [1.23 backport]1mo
I tried creating backports, but there are some merge conflicts. @FiloSottile If you don't mind and have a chance, perhaps it's easier for you to resolve them and send the CLs for review. Thanks.
(…dmitshur commented on log/slog: Custom Handler breaks when using With() and loses parent attributes1mo
CC @jba.dmitshur commented on cmd/go: "cannot access the file because it is being used by another process" flakes on Windows1mo
> AFAICT, there is no tracking issue, just a sea of TestScript testflakes issues, with this specific message showing up about 3 weeks ago in #66337, which has a catch-all pkg == "cmd/go" && test == "…dmitshur commented on golang.org/x/build/cmd/gerritbot: adjust suggestions for wiki repo1mo
Thank you @thepudds! That change is deployed now. Happy New Year to you as well!dmitshur commented on : proxy.golang.org 403 forbidden much like #481071mo
CC @samthanawalla, @findleyr, @hyangah.dmitshur opened an issue github.com/nondebug/gamepad-explorer: Handling vibrationActuator field being null (instead of undefined)1mo
Hello, thanks for offering this page.
The current code has some checks whether a gamepad's vibrationActuator field is undefined.
I've observed cases¹ where a gamepad's vibrationActuator field …dmitshur closed an issue cmd/cgo/internal/testsanitizers: TestTSAN/tsan13 failures1mo
Merging into #67800.dmitshur commented on golang.org/x/net/route: ParseRIB fail to parse utun up InterfaceMessage1mo
CC @neild, @ianlancetaylor.dmitshur commented on golang.org/x/crypto/ssh: `ParseRawPrivateKey` should return `PassphraseMissingError` for `ENCRYPTED PRIVATE KEY`1mo
CC @drakkan, @golang/security.dmitshur commented on cmd/go/internal/modcmd: go.work local module fetching from remote by mistake1mo
Thanks for the report. Are you able to share a go.work and other files needed to trigger the behavior you're seeing, preferably reduced to a minimal set? Also, Go 1.21.1 is a bit old, does this still…dmitshur reviewed +2 on golang.org/x/website/_content/doc/go1.24: add objdump updates2mo
Thanks.dmitshur commented on golang.org/x/open2opaque: Initial commit2mo
(1 comment)
This link can be updated to the final blog post's URL, but it doesn't have to happen in the initial commit.dmitshur opened a change golang.org/x/website/cmd/golangorg: redirect /issue(s) on tip.golang.org2mo
The /issue and its alias /issues subtrees were missed in CL 634915.
Also group related subtrees together.
For golang/go#58484.