[darcs-users] delta-debugging and trackdown --bisect

Jason Dagit dagit at codersbase.com
Tue Oct 6 15:00:48 UTC 2009


On Tue, Oct 6, 2009 at 1:20 AM, Eric Kow <kowey at darcs.net> wrote:

> Thanks for this, Jason,
>
> It's occurred to me that we would be much more likely to make progress
> if we gave ourselves "permission" to implement bisect instead of the
> much cooler Delta Debugging approach (due to my lack of understanding, I
> can only presume it's cooler).
>
> So I've split the ticket:
>  * http://bugs.darcs.net/issue1208 - trackdown --bisect
>   A very *easy* bug that would catch us up nicely
>  * http://bugs.darcs.net/issue1638 - trackdown with delta debugging
>

I think this is wise.

>
> Do we really care about order?
>

I'm not sure.  My reason for suggesting it was 1) to cut down the search
space, 2) I was thinking you'd really only need to check sets of patches
that have existed between invocations of darcs.  Certainly the ordering of
patches shouldn't matter, but the set of patches does matter.


>
> I suppose I ought to just go away and read the paper for a while.
>

It's a good read.  Just understanding the clever tricks they implemented
should give you a bit more insight into doing bisection searches when
debugging manually.

Thanks for creating the bug tickets!
Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20091006/dbfd6a0a/attachment.htm>


More information about the darcs-users mailing list