[darcs-users] GHC segfaulting when building darcss
Simon Marlow
simonmarhaskell at gmail.com
Wed Dec 3 19:30:55 UTC 2008
Ian Lynagh wrote:
> On Wed, Nov 26, 2008 at 08:13:31PM +0100, Reinier Lamers wrote:
>> Hmmm, wasn't Haskell supposed to make segfaults a thing of the past?
>
> You can still get segfaults from at least:
> * compiler bugs
> * runtime system bugs
> * incorrect use of FFI
> * various unsafe Haskell primitives
>
In particular, we had a bug in 6.8.3 and earlier with buffer management
in the code that parses the module header. On Windows this results in
the dreaded "Bad interface file" error that randomly appears, but I
imagine it's possible to get segfaults too. If I recall correctly, the
bug was aggravated by having long comments before the 'module' keyword
in a source file, or lots of {-# OPTIONS #-} pragmas.
Cheers,
Simon
More information about the darcs-users
mailing list