Matthew Pickering

Thank you for bringing this attention to our attention @Kleidukos and @andreabedini !

@ulysses4ever It is something we need to resolve for 3.16. I intend to implement #9115 before then, just lacking time to do so.

> Just to doublecheck, the issue was that `doctest` did not properly handle response files, right? > > Or is there anything else going on? > > * `--with-repl` is...

FWIW, `hadrian` has this behaviour (prefers .hs files over .x) files, which is by design so that we can ship source distributions which don't need `alex`/`happy` to build.

I think the issue happens when you have multiple cursor windows open, opening a new window does not reflect the environment of where the new window is opened but only...

Would it be possible to get a date when this release is going to be ready?

@Mikolaj Not all regressions can be fixed for every release, it doesn't seem like there has been progress on #9777 nor #10290 so they could be bumped to the next...

#10416 is an important fix for the 3.14 release.

Is there perhaps some similar issue where the initialisation events (cost centre definitions) are being posted to the output, and that is interupted by a heap profile event before all...

Another thing to try is a longer profiling interval (`-i10`), does that fix the issue?