(Periodic) Segmentation faults when using regex-compat-0.92

Simon Marlow marlowsd at gmail.com
Tue Jul 13 15:23:19 EDT 2010


It is probably this bug

http://hackage.haskell.org/trac/ghc/ticket/4038

which was present for around a couple of months in HEAD, and appeared in 
GHC 6.12.2 before being fixed in 6.12.3.

If you're on Linux or Windows you can download snapshot distributions of 
GHC from here:

http://www.haskell.org/ghc/dist/current/dist/

but I'd recommend 6.12.3 unless you need HEAD for some other reason.

Cheers,	
	Simon

On 13/07/10 19:27, Kathleen Fisher wrote:
> I am able to repeat the bug using regex 0.93.1 and ghci (version 6.13.20100309).  I've filed a trac report.  In the meantime, any suggestions on what I might do to work around the problem?    Is there a way to upgrade to the most recent HEAD version without going through a complete reinstallation (which takes hours)?
>
> Kathleen
>
>
>
> On Jul 13, 2010, at 4:23 AM, Simon Marlow wrote:
>
>> I tried this here with regex 0.93.1 and GHC build yesterday, and couldn't reproduce it on either 32 or 64-bit x86.
>>
>> Cheers,
>> 	Simon
>>
>> On 13/07/2010 08:13, Simon Peyton-Jones wrote:
>>> Don's right.  If it's reproducible please file a Trac ticket.
>>>
>>> Simon
>>>
>>> | -----Original Message-----
>>> | From: libraries-bounces at haskell.org [mailto:libraries-bounces at haskell.org] On
>>> | Behalf Of Don Stewart
>>> | Sent: 13 July 2010 00:36
>>> | To: Kathleen Fisher
>>> | Cc: TextRegexLazy at personal.mightyreason.com; libraries at haskell.org
>>> | Subject: Re: (Periodic) Segmentation faults when using regex-compat-0.92
>>> |
>>> | kathleen.fisher:
>>> |>   I am getting periodic (although frequent) segmentation faults when loading
>>> | the following code into ghci (version 6.13.20100309) and then asking for the
>>> | value of re_results1:
>>> |>
>>> |>   >   module Examples.RegExpTest where
>>> |>   >
>>> |>   >   import qualified Text.Regex as RE
>>> |>   >
>>> |>   >   re = RE.mkRegexWithOpts "^a+" True True
>>> |>   >   re_results1 = RE.matchRegexAll re "aaaab"
>>> |>
>>> |>   Is this a known problem?  Are there known work arounds?
>>> |
>>> | That sounds very scary. There should never be a segfault in code like
>>> | this. It may point to a compiler bug in the GHC head.
>>> |
>>> | I wonder if you can upgrade to regex-compat 0.93.1 ?
>>> |
>>> |     http://hackage.haskell.org/package/regex-compat
>>> |
>>> | Forwarding to the maintainer, Chris Kuklewicz.
>>> |
>>> | -- Don
>>> |
>>> | _______________________________________________
>>> | Libraries mailing list
>>> | Libraries at haskell.org
>>> | http://www.haskell.org/mailman/listinfo/libraries
>>>
>>> _______________________________________________
>>> Libraries mailing list
>>> Libraries at haskell.org
>>> http://www.haskell.org/mailman/listinfo/libraries
>>
>



More information about the Libraries mailing list