Stg-lint and bootstrapping (fails?)

[Ben Gamari]

Gabor Greif <ggreif at gmail.com> writes:

> Hi Ben,
>
> thanks for also looking into the issue! I am in the process of
> debugging a miscompilation related to #13861 (self-inflicted,
> hahahar), but without a working linter I am pretty much toast... I'll
> wait a bit more until your fixed hit master.
>
If it helps I can give you a conservative "fix" which should at least
allow the linter to pass valid code (at the expense of missing some bad
programs). Ultimately this may be what we are focused to accept anyways.

I'll discuss this with Simon next week.

> Unrelated to linting, but seemingly related to "-g" I get a ton of
> warnings like this:
>
> WARNING: file compiler/simplCore/OccurAnal.hs, line 2695
>   OccurAnal failed to rediscover join point(s): [go58_akKa]
>
> Is this perhaps covered by your fixes?
>
I'm afraid not; this is the simplifier complaining that it lost track of
a join point which should be solely an optimization.

Cheers,

- Ben
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 487 bytes
Desc: not available
URL: <http://mail.haskell.org/pipermail/ghc-devs/attachments/20170818/1893a683/attachment.sig>