Patch 907 resolve issue2225: add an option to obliterate all pat...

Title resolve issue2225: add an option to obliterate all pat...
Superseder Nosy List ganesh, owst
Related Issues
Status accepted Assigned To owst

Created on 2012-08-19.01:29:25 by owst, last changed 2014-11-15.12:55:31 by owst.

File name Status Uploaded Type Edit Remove
patch-preview.txt owst, 2012-08-19.01:29:25 text/x-darcs-patch
resolve-issue2225_-add-an-option-to-obliterate-all-patches-not-in-remote-repo.dpatch owst, 2012-08-19.01:29:25 application/x-darcs-patch
unnamed owst, 2012-08-19.01:29:25
See mailing list archives for discussion on individual patches.
msg15989 (view) Author: owst Date: 2012-08-19.01:29:25
This patch is a first stab at this. It'd be nice to be able to not require a
URL, by taking the default remote as the default.

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

Sun Aug 19 02:29:26 BST 2012  Owen Stephens <darcs@owenstephens.co.uk>
  * resolve issue2225: add an option to obliterate all patches not in remote repo
msg16266 (view) Author: galbolle Date: 2012-10-31.09:48:21
Just one bit of grumpy old man: wouldn't it make sense to have the
not-in-repo integrated to --match and available to all commands? For
instance, it could probably help with rebase suspend or amend-record.
This implementation also seems to (silently?) ignore --match if
--not-in-repo is given, which is not good.
msg16268 (view) Author: owst Date: 2012-10-31.10:54:18
Yes, almost certainly, re: --match.

I hadn't thought of the applicability to rebase/amend, good ideas, and
persuasive for making it part of --match.

I'll take a look at some point, and try and integrate it in a nice way -
including the ignoring of --match flags as *is* currently (stupidly) the

Thanks Florent!
msg16364 (view) Author: owst Date: 2012-11-12.01:52:31
I'm reluctantly pushing this to reviewed since 2.9.4 and 2.9.5 depend on
it. I will follow-up later.
msg16505 (view) Author: ganesh Date: 2012-12-29.21:13:55
I think it would be hard to make this an option in --match (a "primitive matcher") because the 
efficient implementation depends on being able to do findCommonAndUncommon in a single pass 
(and thus take account of clean tags etc), whereas --match would allow for things like --match 
"(in-repo foo && hunk wibble) || (!in-repo bar && hunk wobble)" which certainly can't easily 
be implemented that way.

I'm not sure if it would make sense to integrate it with Darcs.Patch.Match in the way various 
other patch-selection options are.

Certainly rebase suspend would benefit from it.
msg17166 (view) Author: ganesh Date: 2014-02-08.12:39:26
Ping, re following up later :-)
msg17729 (view) Author: ganesh Date: 2014-10-31.07:50:36
*reprod* - do you still plan to do any more with this, or should we 
just leave things as they are?

One thing I keep wanting is for it to respect the same default repo as 
push/pull do, so I don't have to retype it.
msg17799 (view) Author: owst Date: 2014-11-15.12:55:31
Two weeks late after prodding, but I've just sent a new follow-up patch
to this that allows multiple repos and defaults to the push/pull default
repo. Sorry about the silence and delay!
Date User Action Args
2012-08-19 01:29:25owstcreate
2012-08-26 01:26:46ganeshsetstatus: needs-screening -> needs-review
2012-10-31 09:48:22galbollesetmessages: + msg16266
2012-10-31 09:48:38galbollesetstatus: needs-review -> followup-requested
assignedto: owst
2012-10-31 10:54:18owstsetmessages: + msg16268
2012-11-12 01:52:31owstsetmessages: + msg16364
2012-12-29 21:13:55ganeshsetnosy: + ganesh
messages: + msg16505
2014-02-08 12:39:26ganeshsetmessages: + msg17166
2014-10-31 07:50:36ganeshsetmessages: + msg17729
2014-11-04 23:03:06ganeshsetstatus: followup-requested -> accepted
2014-11-15 12:55:31owstsetmessages: + msg17799