darcs

Issue 831 wish: consistent ordering of patches

Title wish: consistent ordering of patches
Priority wishlist Status waiting-for
Milestone Resolved in
Superseder minimal context function
View: 1514
Nosy List darcs-devel, dleverton, dmitry.kurochkin, kowey, markstos, thorkilnaur, tommy, zooko
Assigned To
Topics

Created on 2008-05-02.18:54:22 by zooko, last changed 2010-03-23.23:45:49 by kowey.

Messages
msg4494 (view) Author: zooko Date: 2008-05-02.18:54:20
I wish that a command (possibly just "darcs optimize --reorder") would sort the
patches into some consistent order, such that if I run that command one one
repository, and I run that command on another repository which has exactly the
same set of patches in it, then afterward the two repositories will have exactly
the same order.

This is useful for trac, which assigns a sequence number to each patch, e.g.
http://allmydata.org/trac/darcs-2 , and for the versioning scheme that we use
for allmydata.org "Tahoe" project, where the intermediate releases are numbered
by how many patches were in them since the most recent tag:

http://allmydata.org/source/tahoe/tarballs/

It would also be useful for a new versioning scheme that I am thinking of adding
to Tahoe -- take the hash of the output of the "darcs changes --context" and
include that in the version number, so that if you see two versions with the
same hash code then you know that the two were built from the same version of
the source.

Finally, this consistent sort order could be useful for two programmers who are
"comparing" notes and reading through their changes.
msg5451 (view) Author: markstos Date: 2008-08-13.01:13:13
Zooko,

It's reasonable idea. It seems work discussing on the mailing list to gather
more feedback.

Would this allow you to 'obliterate' a patch that is being tracked with 'Trac'.
I can't see that it would. Perhaps that's a seperate issue, and one that could
be solved on the Trac side.
msg5454 (view) Author: zooko Date: 2008-08-13.01:41:44
Hey Mark Stosberg -- welcome back!  :-)
msg8367 (view) Author: kowey Date: 2009-08-23.01:00:23
Unfortunately, I don't think optimize --redorder  will do what you want.  

I believe this would just be an extension of the work needed for minimal
contexts so I'm marking this deferred until we get there.

I suppose this could then work its way into the reorder code so that patches
also get reordered according to scheme.
History
Date User Action Args
2008-05-02 18:54:22zookocreate
2008-08-13 01:13:17markstossetstatus: unread -> unknown
nosy: + markstos
messages: + msg5451
2008-08-13 01:41:47zookosetnosy: tommy, beschmi, markstos, zooko, dagit
messages: + msg5454
2009-01-27 23:34:23dlevertonsetnosy: + dmitry.kurochkin, dleverton, simon, kowey, thorkilnaur
2009-08-06 21:03:55adminsetnosy: - beschmi
2009-08-11 00:12:30adminsetnosy: - dagit
2009-08-23 01:00:25koweysetstatus: unknown -> deferred
nosy: tommy, kowey, markstos, zooko, simon, thorkilnaur, dmitry.kurochkin, dleverton
superseder: + minimal context function
messages: + msg8367
2009-08-25 17:40:04adminsetnosy: + darcs-devel, - simon
2009-08-27 14:29:57adminsetnosy: tommy, kowey, markstos, darcs-devel, zooko, thorkilnaur, dmitry.kurochkin, dleverton
2010-03-23 23:45:49koweysetstatus: deferred -> waiting-for