1 Tag rename unfinished business
2 (things resulting from searching for `tag' where searching for
3 `debiantag' wouldn't find them)
6 implies ability to sign new tag (this is already there
7 so simply advertising version 4 will work)
8 needs also new ability to sign old-format tag in split
13 - other tag should be [dgit Maintainer-Tag quilt=$quilt-mode]
15 - in dopush (?), we need to sort out synthetic fast forard
16 this should not depend entirely on quilt mode
17 but will always be needed if quiltmode_splitbrain
18 * On push, if ff fixup mode enabled, check debian/changelog,
19 and maybe make fake merge, maybe in dgit view, maybe
22 - there is some confusion between $split_brain (did we actually
23 generate any commits) and quiltmode_splitbrain (might we have done,
24 including in previous pushes); search for $split_brain and check
25 that use sites are correct
29 - start repo with old tag, push with new tag, run
30 "old" dgit to push again with new tag, push again with
32 - rejecting push of same version with one tag when then the other
34 Test that, where appropriate, maint view tags are pushed to dgit git repo
37 * Provide --gpbpq[=:] options
39 * Document --quilt=unapplied|gbp|dpm
41 * Mention new --quilt=* in debian/changelog
43 * Mention in-archive copy fix in debian/changelog
48 * IN ARCHIVE COPY Consider and implement fast-forward fixup
50 - On fetch, check if specified git hash is ff from
51 existing dgit branch, the same way we do for constructed
52 dscs. The .dsc tells us the git tree but not necessarily
57 complicated there are many branches
58 dgit-repos: refs/dgit/suite
60 local; remotes/dgit/dgit/suite
63 local: dgit/suite (not relevant to fetch)
65 archive dsc must always be regarded
66 as ff from (or same as) dgit-repos
67 err depending on version number ?
69 this code is what needs changing
71 fail "git head (".lrref()."=$lastpush_hash) is not a ".
72 "descendant of archive's .dsc hash ($dsc_hash)";
74 it just needs to generate the fake merge and set $hash
75 the archive skew stuff is all fine with this
77 xxx actually, no, we need to stop from squashing our
80 * Add in-archive-move test case