Yeah, I should have said suite/package, not suite/file, because of course packages can have many test files. Internally at Google, the go test output is transformed into Sponge XML, a superset of J...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "golang.org/x/tools/internal/modindex" && test == "TestIncremental"
<details><summary>2024-10-24 15:50 x_tools-gotip-s...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "golang.org/x/tools/cmd/stringer" && test ~ `TestEndToEnd`
<details><summary>2024-10-24 15:50 x_tools-gotip-solaris-am...
> note that we can't change the `Channel` interface definition for backwards compatibility reasons.
Does that freeze include adding a method?
As I don't want to change any existing function signa...
I updated CL 600296 to cover this, implemented basically https://github.com/golang/go/issues/68587#issuecomment-2304877746. If the stack moved, we use a new bounds of [SP, SP+32KB), which I think i...
@xzbdmw I think the tricky thing is defining the name in the lexical scope. Right now, gopls looks in the surrounding `types.Scope` to find available names. If the name is undefined due to the inva...
This issue is extracted from #69980, where the request is to include undeclared (but used) names in lexical completion results.
We could do this by extracting such names from type errors (see also...
I imagine it would just use a strings.Builder and opt into whatever the default buffer there is. Maybe there could be optimization to use a slice for sequences shorter than N.
Adding support for properties-in-output to go-junit-report seems like an excellent path to generating data on the general usefulness of properties.
> Even if such support is added to go-junit-re...
Do you imagine that it would collect all the items from the sequence first to figure out the total length of the final string before concatenating them?
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "cmd/cgo/internal/testfortran" && test == "TestFortran"
<details><summary>2024-10-24 16:55 gotip-solaris-amd64 go@fb2a...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "golang.org/x/tools/internal/modindex" && test == "TestIncrementalNope"
<details><summary>2024-10-24 15:50 x_tools-got...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "cmd/cgo/internal/testerrors" && test == "TestReportsTypeErrors"
<details><summary>2024-10-24 16:55 gotip-solaris-amd6...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "golang.org/x/tools/internal/modindex" && test == "TestIncremental"
<details><summary>2024-10-24 15:50 x_tools-gotip-n...
Found new dashboard test flakes for:
#!watchflakes
default <- pkg == "golang.org/x/tools/cmd/stringer" && test ~ `TestEndToEnd`
<details><summary>2024-10-24 15:50 x_tools-gotip-solaris-am...
Found new dashboard test flakes for:
#!watchflakes
post <- (pkg ~ `^cmd/go` || pkg ~ `^golang.org/x/mod`) && (
`Failed to connect to github.com port 443: Connection refused` ||
`...
Huzzah! `make.bash` works now. https://ci.chromium.org/ui/p/golang/builders/ci-workers.shadow/gotip-linux-arm64-clang15-build_go/b8733166898151394753/infra
I am experimenting with the implementation, and feel this might be a better semantic:
For the loop body of testing.B.Loop, no inlining will be performed. (inlining could still happen inside the ...
### Go version
go version devel go1.24-4646556ba4 Thu Oct 24 12:32:56 2024 +0000 solaris/amd64
### Output of `go env` in your module/workspace:
```shell
GO111MODULE=''
GOARCH='amd64'
GOAUTH='n...
### Proposal Details
#61901 added SplitSeq and friends to bytes, but nothing for joining. Right now you have to do slices.Collect first or use a bytes.Buffer/strings.Builder, which feels unfortuna...