[darcs-users] forensics case study (isolating an 'Irrefutable pattern failed' error)

[Eric Kow]

On Fri, Feb 20, 2009 at 12:27:23 +0100, Christian Kellermann wrote:
> So how should we proceed with the original bug? I can go on not
> using the patch bundle but to use the common + pulled repo as the
> initial state for minimalising the history?

I'm afraid I'm now a little bit lost (despite my attempts at re-reading
the thread).

We have three repositories, common mainline and branch.  We can call the
patch bundles C-M (patches starting from common and ending in mainline)
C-B (common => branch) and M-B

You're saying that
 
 mainline: apply M-B.dpatch    # ERROR! irrefutable pattern

 common : pull -a ../mainline  # no conflicts
          apply M-B.dpatch     # ERROR! irrefutable pattern

 common : apply C-B.dpatch     # conflicts

Is that correct?  What I would be interested to see is something like
this:

  common: apply C-M # no conflicts
          apply M-B # ERROR! irrefutable pattern

Is that possible?  The apply C-M should take the place of the pulling
from mainline...

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20090227/5a1192a8/attachment.pgp>