darcs

Patch 1843 bugfix in V3.commuteNoConflicts

Title bugfix in V3.commuteNoConflicts
Superseder Nosy List bf
Related Issues
Status needs-review Assigned To
Milestone

Created on 2019-06-24.15:30:49 by bf, last changed 2019-06-28.11:57:50 by bf.

Files
File name Status Uploaded Type Edit Remove
bugfix-in-v3_commutenoconflicts.dpatch bf, 2019-06-25.09:29:11 application/x-darcs-patch
patch-preview.txt bf, 2019-06-25.09:29:11 text/x-darcs-patch
unnamed bf, 2019-06-25.09:29:11 text/plain
See mailing list archives for discussion on individual patches.
Messages
msg20871 (view) Author: bf Date: 2019-06-24.15:30:49
See the mailing list thread "new test case generator".

1 patch for repository http://darcs.net/screened:

patch 45dea0474f011b82e354a7554c064351d940e6ff
Author: Ben Franksen <ben.franksen@online.de>
Date:   Mon Jun 24 10:09:20 CEST 2019
  * bugfix in V3.commuteNoConflicts
  
  I made a logical error when I derived the mixed Conflictor/Rotcilfnoc cases
  for commuteNoConflicts from the Conflictor/Conflictor case. This could lead
  to the creation of invalid Conflictors during an unconflicted merge of two
  Conflictors. Specifically, we could create Conflictors that undo prims that
  are already undone by a previous Conflictor.
Attachments
msg20872 (view) Author: bf Date: 2019-06-24.19:56:36
Please don't screen yet. I think I have made yet another mistake and
"fixed" one case too many. And I think the reason this doesn't lead to
failures in the tests is that mergeNoConflicts always commutes "invert
lhs:> rhs", so we don't run into the Conflictor:\/:Rotcilfnoc case. I
will try to find a way to make sure we test that case, too. I should
also instrument darcs-test to get coverage information.
msg20873 (view) Author: bf Date: 2019-06-25.08:02:54
This obsoletes the patch I sent earlier.

1 patch for repository http://darcs.net/screened:

patch 230fd522fb61feaf4ae4f50d0f0c55beee788bf8
Author: Ben Franksen <ben.franksen@online.de>
Date:   Mon Jun 24 10:09:20 CEST 2019
  * bugfix in V3.commuteNoConflicts
  
  I made a logical error when I derived the mixed Rotcilfnoc/Conflictor case
  for commuteNoConflicts from the Conflictor/Conflictor case. This could lead
  to the creation of invalid Conflictors during an unconflicted merge of two
  Conflictors. Specifically, we could create Conflictors that undo prims that
  are already undone by a previous Conflictor.
  The only reason we need this case at all is because mergeNoConflicts
  triggers it when applied to a pair of Conflictors. This means that the
  mirror case Conflictor/Rotcilfnoc is not needed, since we currently have no
  need for an inverse operation to mergeNoConflicts. So this case is now
  commented out and replaced with an error call.
Attachments
msg20874 (view) Author: bf Date: 2019-06-25.09:29:11
...remembering that leaving obsolete code in comments is a bad idea. We can
always roll back or view the patch to see what was there.

1 patch for repository http://darcs.net/screened:

patch 3e205eb49da2c391498d4c8e922a1e4f675b6622
Author: Ben Franksen <ben.franksen@online.de>
Date:   Mon Jun 24 10:09:20 CEST 2019
  * bugfix in V3.commuteNoConflicts
  
  I made a logical error when I derived the mixed Rotcilfnoc/Conflictor case
  for commuteNoConflicts from the Conflictor/Conflictor case. This could lead
  to the creation of invalid Conflictors during an unconflicted merge of two
  Conflictors. Specifically, we could create Conflictors that undo prims that
  are already undone by a previous Conflictor.
  The only reason we need this case at all is because mergeNoConflicts
  triggers it when applied to a pair of Conflictors. This means that the
  mirror case Conflictor/Rotcilfnoc is not needed, since we currently have no
  need for an inverse operation to mergeNoConflicts. So this case is now
  replaced with an error call.
Attachments
History
Date User Action Args
2019-06-24 15:30:49bfcreate
2019-06-24 19:56:37bfsetmessages: + msg20872
2019-06-25 08:02:54bfsetfiles: + patch-preview.txt, bugfix-in-v3_commutenoconflicts.dpatch, unnamed
messages: + msg20873
2019-06-25 09:29:11bfsetfiles: + patch-preview.txt, bugfix-in-v3_commutenoconflicts.dpatch, unnamed
messages: + msg20874
2019-06-25 09:34:04bfsetfiles: - bugfix-in-v3_commutenoconflicts.dpatch
2019-06-25 09:34:08bfsetfiles: - bugfix-in-v3_commutenoconflicts.dpatch
2019-06-25 09:34:13bfsetfiles: - patch-preview.txt
2019-06-25 09:34:16bfsetfiles: - patch-preview.txt
2019-06-25 09:34:19bfsetfiles: - unnamed
2019-06-25 09:34:20bfsetfiles: - unnamed
2019-06-28 11:57:50bfsetstatus: needs-screening -> needs-review