Patch 1820 accept issue2609: conflicts in unrevert bundle

Title accept issue2609: conflicts in unrevert bundle
Superseder Nosy List bf
Related Issues
Status accepted Assigned To

Created on 2019-06-13.12:05:27 by bf, last changed 2019-07-28.19:38:33 by ganesh.

File name Status Uploaded Type Edit Remove
accept-issue2609_-conflicts-in-unrevert-bundle.dpatch bf, 2019-06-13.12:05:27 application/x-darcs-patch
issue2609-is-not-a-bug-but-expected-behavior.dpatch bf, 2019-06-16.10:21:30 application/x-darcs-patch
patch-preview.txt bf, 2019-06-13.12:05:27 text/x-darcs-patch
patch-preview.txt bf, 2019-06-16.10:21:30 text/x-darcs-patch
unnamed bf, 2019-06-13.12:05:27 text/plain
unnamed bf, 2019-06-16.10:21:30 text/plain
See mailing list archives for discussion on individual patches.
msg20773 (view) Author: bf Date: 2019-06-13.12:05:27
I would like to hear opinions about this: is this a bug or a feature? do we
want the unrevert bundle contain conflicts?

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

patch fe565b18beaf9169a28b8c49464904a0f27e0f5a
Author: Ben Franksen <ben.franksen@online.de>
Date:   Thu Nov 15 15:18:07 CET 2018
  * accept issue2609: conflicts in unrevert bundle
msg20792 (view) Author: ganesh Date: 2019-06-14.12:18:55
Feels like a feature to me, but I'm not certain.
msg20827 (view) Author: bf Date: 2019-06-15.10:22:01
I find it interesting that this test fails with darcs-3 patch format, 
too, even though the underlying scenario (issue2605) no longer does.
msg20840 (view) Author: bf Date: 2019-06-15.19:32:32
I should read my own comments. Nothing to do with issue2605, see 
msg20465. Also conflict is not in the unrevert bundle but appears when 
we apply it, see msg20839. So I guess this is a false alarm and I 
should rollback.
msg20841 (view) Author: ganesh Date: 2019-06-15.21:28:59
My view is that unrevert conflicting with recorded is not
very surprising, and that unrevert containing conflicts may or may
not be valid.

Whenever either scenario can arise, presumably the alternative is 
just to throw it away. So I'd be inclined to let it happen.

On the other hand if the only way you can cause unrevert to contain
conflicts is with the duplicate patch hack like in msg20460, then
I almost don't care what happens because I think that hack is 
already taking us into very messy territory and this is the least
of the problems.
msg20842 (view) Author: bf Date: 2019-06-16.10:21:30
There is no connection with duplicate patches, i just stumbled over this
behavior when experimenting with them. The conflict is not /in/ the unrevert
bundle, it happens due to unrevert conflicting with recorded. This is
normal, expected behavior. Instead of rolling back, I thought it better to
turn it into a succeeding test, documenting the expected behavior.

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

patch 604ff48230dd0c6ef74216f3d9c8ea856a4436c4
Author: Ben Franksen <ben.franksen@online.de>
Date:   Sun Jun 16 12:16:32 CEST 2019
  * issue2609 is not a bug but expected behavior
Date User Action Args
2019-06-13 12:05:27bfcreate
2019-06-14 12:18:55ganeshsetmessages: + msg20792
2019-06-15 10:22:01bfsetstatus: needs-screening -> needs-review
messages: + msg20827
2019-06-15 19:32:32bfsetmessages: + msg20840
2019-06-15 21:29:00ganeshsetmessages: + msg20841
2019-06-16 10:21:30bfsetfiles: + patch-preview.txt, issue2609-is-not-a-bug-but-expected-behavior.dpatch, unnamed
messages: + msg20842
2019-07-27 18:20:13ganeshsetstatus: needs-review -> accepted-pending-tests
2019-07-28 19:38:33ganeshsetstatus: accepted-pending-tests -> accepted