<div dir="auto">Don’t forget ghc 9 is already out! :)</div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Feb 15, 2021 at 2:10 PM Troels Henriksen <<a href="mailto:athas@sigkill.dk">athas@sigkill.dk</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">It is very likely that issue 17386 is the issue. With<br>
<br>
{-# OPTIONS_GHC -Wno-overlapping-patterns -Wno-incomplete-patterns<br>
-Wno-incomplete-uni-patterns -Wno-incomplete-record-updates #-}<br>
<br>
my module(s) compile very quickly. I'll wait and see if GHC 9 does<br>
better before I try to create a smaller case (and now I at least have a<br>
workaround).<br>
<br>
Sebastian Graf <<a href="mailto:sgraf1337@gmail.com" target="_blank">sgraf1337@gmail.com</a>> writes:<br>
<br>
> Hi,<br>
><br>
> I'm not sure I see all the context of the conversation, but it is entirely<br>
> possible that code with many local constraints regresses the pattern-match<br>
> checker (which is accounted to Desugaring in the profile emitted by -v2),<br>
> I'm afraid. That simply has to do with the fact that we now actually care<br>
> about them, previously they were mostly discarded.<br>
><br>
> I'd be glad if you submitted a relatively isolated reproducer of what is<br>
> fast with 8.8 and slow with 8.10 (even better 9.0).<br>
> I hope that things have improved since we fixed<br>
> <a href="https://gitlab.haskell.org/ghc/ghc/-/issues/17836" rel="noreferrer" target="_blank">https://gitlab.haskell.org/ghc/ghc/-/issues/17836</a>, which is part of 9.0 but<br>
> not of 8.10.<br>
><br>
> Cheers,<br>
> Sebastian<br>
><br>
> Am Mo., 15. Feb. 2021 um 19:04 Uhr schrieb Troels Henriksen <<br>
> <a href="mailto:athas@sigkill.dk" target="_blank">athas@sigkill.dk</a>>:<br>
><br>
>> Carter Schonwald <<a href="mailto:carter.schonwald@gmail.com" target="_blank">carter.schonwald@gmail.com</a>> writes:<br>
>><br>
>> > Ccing ghc devs since that’s a better forum perhaps<br>
>> > Crazy theory:<br>
>> ><br>
>> > this is a regression due the the partial changes to pattern matching<br>
>> > coverage checking in 8.10 that finished / landed in ghc 9<br>
>> ><br>
>> > Why:<br>
>> > Desugaring is when pattern/ case statement translation happens I think?<br>
>> > And the only obvious “big thing” is that you have some huge , albeit sane<br>
>> > for a compiler, pattern matching<br>
>> ><br>
>> > I’d first check if the new ghc 9 release doesn’t have that regression in<br>
>> > build time that you experienced. And if it does file a ticket.<br>
>> ><br>
>> > I may be totally wrong, but that seems like a decent likelihood !<br>
>><br>
>> You may be right! Another module that regressed is also mainly<br>
>> characterised by large-but-not-insane case expressions:<br>
>><br>
>> <a href="https://github.com/diku-dk/futhark/blob/d0839412bdd11884d75a1494dd5de5191833f39e/src/Futhark/Optimise/Simplify/Rules.hs" rel="noreferrer" target="_blank">https://github.com/diku-dk/futhark/blob/d0839412bdd11884d75a1494dd5de5191833f39e/src/Futhark/Optimise/Simplify/Rules.hs</a><br>
>><br>
>> I'll try to split these modules up a little bit (I should have done so a<br>
>> while ago anyway) and maybe that will make the picture even clearer.<br>
>><br>
>> --<br>
>> \ Troels<br>
>> /\ Henriksen<br>
>> _______________________________________________<br>
>> ghc-devs mailing list<br>
>> <a href="mailto:ghc-devs@haskell.org" target="_blank">ghc-devs@haskell.org</a><br>
>> <a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs" rel="noreferrer" target="_blank">http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs</a><br>
>><br>
<br>
-- <br>
\ Troels<br>
/\ Henriksen<br>
</blockquote></div></div>