David Roundy <[EMAIL PROTECTED]> writes:

> On Tue, Dec 09, 2008 at 09:22:34AM +0100, Petr Rockai wrote:
>> Hi,
>> 
>> for most of the patch, I don't have an opinion (ie. I don't quite see what's
>> going on there, the test flags are sort of opaque). However, I need to ask, 
>> are
>> you sure about those exitWith's replacing the return? I understand that will
>> circumvent the cleanup replayRepository is doing... I'm not sure the 
>> `finally`
>> is enough to pre-empt exitWith. Maybe it is and then it's all fine...
>
> Yes, the key is to use the exception handling that comes in
> Control.Exception, rather than the exception handling that comes in
> System.IO.  The former catches things like asynchronous exceptions and
> exitWith, while the latter doesn't.
Thanks for the explanation. Therefore, I have no further objections against
these patches. Repository.Repair indeed uses Control.Exception's finally.

However, this (unfortunately) doesn't mean that I have reviewed the patch, just
skimmed it. A separate review by someone else might be in place. Btw., do you
think it would be possible to describe the fixed bug in such a way that someone
(I probably volunteer) could write a testcase for it?

Yours,
   Petr.

-- 
Peter Rockai | me()mornfall!net | prockai()redhat!com
 http://blog.mornfall.net | http://web.mornfall.net

"In My Egotistical Opinion, most people's C programs should be
 indented six feet downward and covered with dirt."
     -- Blair P. Houghton on the subject of C program indentation
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to