At the moment, hashable-1.2 generates consistent segfaults on some machines, on both Linux and Windows (tibbe/hashable#49). It'd be nice if this version could be temporarily blacklisted with a hashable < 1.2 constraint so that users weren't required to do it manually. The Yesod guys have already done it.
hashable < 1.2
edit: The issue has been fixed in HEAD. Still may be worth it to blacklist hashable >= 188.8.131.52 && < 184.108.40.206.
I'd prefer to wait a few days to see if @tibbe will fix the underlying problem. I just got back from vacation and don't see having enough cycles to put out a new release in the next few days anyways, maybe @mightybyte has time.
I've released 220.127.116.11 which should address some of the issues. I'm working on the hashable issue the upcoming week.
The segfault still happens with 18.104.22.168.
I believe there were a few different ones. A few are fixed by 22.214.171.124. Still some more to go (tracked on the bug tracker).
@tibbe I encountered two bugs:
Btw, would blacklisting hashable-1.2 as an afterthought work at all? I mean, wouldn't the cabal-install solver then just backtrack to to the previous version 0.9.2.N-1 which wasn't in conflict with hashable-1.2?
The issue has been fixed in current HEAD (what will be 126.96.36.199, I believe) - see tibbe/hashable@e64484d and discussion in tibbe/hashable#49. At least on my machine, I don't get segfaults anymore. Still may be worth it to blacklist hashable >= 188.8.131.52 && < 184.108.40.206.
(just small note, 220.127.116.11 containing the fix is already released)
Unfortunately, 18.104.22.168 contains another bug - see tibbe/hashable#60. 22.214.171.124, which contains the fix, was released yesterday.
Blacklist bad versions of hashable (closes #164)