Alan Donovan

I think we could probably delete $GOROOT/src/go/internal/gcimporter/iimport.go today. It is only used as the implementation of `go/importer.ForCompiler("gc")`, aka importer.Default(), when it encounters an 'indexed' file, but since it only reads...

@timothy-king I've sent a CL for the deletion. In any case I don't think the concern has any bearing on this proposal.

> Have all remaining concerns about this proposal been addressed? The `iimport` implementation in GOROOT that @timothy-king mentioned has now been deleted. > @aarzilli If for some reason you are...

Excited to use this! A couple of suggestions: - The wording should be "Add" (since there may already be one), should avoid "unit", and should name the symbol. No "Source:"...

All good, though in item 7 I suggest splitting the two cases as it leads to better error messages. (In Google readability reviews people sometimes say "but it's longer", but...

Test timeout after 10m, vast number of goroutines (>1000), many in lsprpc.ServeStream (why so many?), lots of still tests running. A call to ResolveCodeAction is blocked on the client. Not...

It appears that around 577 conn.run calls are active at once, plus (only?!) 287 streamServer.ServeStream calls. Each one of them looks to be doing something reasonable but none is making...

Buried in the JSON log and almost impossible to find is this, implicating TestMCPCommandHTTP: ``` {"Time":"2025-07-08T17:16:36.987572277Z","Action":"output","Package":"golang.org/x/tools/gopls/internal/cmd","Test":"TestMCPCommandStdio","Output":"=== RUN TestMCPCommandStdio\n"} {"Time":"2025-07-08T17:16:37.059941617Z","Action":"output","Package":"golang.org/x/tools/gopls/internal/filecache","Output":"ok \tgolang.org/x/tools/gopls/internal/filecache\t2.371s\n"} {"Time":"2025-07-08T17:16:37.138010956Z","Action":"pass","Package":"golang.org/x/tools/gopls/internal/filecache","Elapsed":2.449} {"Time":"2025-07-08T17:16:39.744908694Z","Action":"output","Package":"golang.org/x/tools/gopls/internal/cmd","Test":"TestMCPCommandStdio","Output":"--- PASS: TestMCPCommandStdio (2.76s)\n"} {"Time":"2025-07-08T17:16:39.744946054Z","Action":"pass","Package":"golang.org/x/tools/gopls/internal/cmd","Test":"TestMCPCommandStdio","Elapsed":2.76} {"Time":"2025-07-08T17:16:39.744958454Z","Action":"run","Package":"golang.org/x/tools/gopls/internal/cmd","Test":"TestMCPCommandHTTP"} {"Time":"2025-07-08T17:16:39.744961614Z","Action":"output","Package":"golang.org/x/tools/gopls/internal/cmd","Test":"TestMCPCommandHTTP","Output":"=== RUN TestMCPCommandHTTP\n"}...

I have filed https://github.com/golang/go/issues/74516 for the builder UI regression.

> Hi Alan, sorry I miss this issue. No worries, it's very hard to see! > Do you think there is a relationship between this test and the callhierarchy marker...