Simon Marlow | 7 Jul 13:01 2009
Picon

Re: failure in issue1248.sh

On 06/07/2009 17:10, Petr Rockai wrote:
> Hi,
>
> Ganesh Sittampalam<ganesh <at> earth.li>  writes:
>> darcs: _darcs/patches/pending.new-0: openBinaryFile: resource busy (file is
>> locked)
>>
>> It seems to be triggered by the following patch:
>>
>> Wed Jun  3 21:37:02 BST 2009  Petr Rockai<me <at> mornfall.net>
>>    * Fix index invalidation in the move command.
>>    Ignore-this: e2fa46552a1ac28aea587c96c0932298
>
> I have managed to reproduce this a few times in various tests. It seems that
> issue1248.sh is just the one that fails "most often", with a number of other
> tests failing from time to time. I have seen issue1248.sh pass with same binary
> that usually fails, too. All this with stock Debian GHC 6.8.2 (stable). I have
> never seen the problem happen with GHC 6.10.3. Moreover, the mentioned patch
> does not seem to change anything about this behaviour for me.
>
> The code that is failing seems to reside in Darcs.Lock, likely the atomic write
> file operations: writeAtomicFilePS and its helpers. I haven't had more time to
> isolate the problem so far. I'll try to resolve this in time for beta 2 though.

Could be this bug:

http://hackage.haskell.org/trac/ghc/ticket/2122

fix in GHC 6.8.3.

Cheers,
	Simon

Gmane