1 What comes on my mind but it isn't able to hold it all. Unsorted.
9 (We have crash courses, one for version control newbies is missing -
10 it should be included in the source as well, though.)
12 * Show only first 12 (or so) nibbles of the hashes everywhere
13 Even this might be too much, but more than this is really useless
14 for anyone remotely human. And it's less scary, too.
17 * Make one-time merges easier by enabling cg-update to take
18 URIs directly, not only head names.
20 * Support for mass-cloning all remote refs at once
21 * Requires first multi-origin support (if origin is a dir, fall
22 back to origin/master).
23 * Interface: cg-clone -a
24 * Implement this using refs/remote/ so that the branches won't
25 get mixed together and we can trace new/disappearing branches
26 in the parent repository sanely
28 * Support for remotes/
29 More powerful 'n stuff than branches/, ya know.
30 LIKELY NOT - parsing those would be just annoying, we will probably
31 just dump directly to supporting [remote] in .git/config. There will
32 need to be a conversion script though.
34 * Push new tags automagically
35 (With the same logic as when fetching - push tags which reference
36 objects that are already pushed.)
37 (First there really needs to be a way to have private tags. I think
38 just tag names starting by a dot would do.)
41 * Handle all possible conflicts
42 * Make sure the user resolved the conflicts
43 (if possible without requiring him to manually indicate that)
44 * Support for various GIT's "merge strategies"
45 * Somehow show the list of commits getting merged
48 * cg-resolve: resolving swiss-army knife, see msgid:
49 <20060924171246.GY13132@pasky.or.cz>
52 cg-mkpatch was overlooked for too long, while it could do a huge
53 number of cool things (after some cleanups, and probably factoring
54 out the common parts of cg-mkpatch, cg-log and cg-diff):
55 Frontending git-request-fetch, interfacing with email directly...
57 * cg-commit enhancements
58 * Force the user to confirm his credentials for the first time
59 * Possibly verify the patch being committed against
60 whitespace errors and stuff; hooks are good for this too,
61 but I think it's good to have internal support for the
62 basic stuff. (Get inspired in git-commit.)
63 * Call git-fmt-merge-msg when committing a merge
65 * cg-shelf - shelve changes temporarily
66 This saves your current uncommitted changes, removes them from the
67 tree, opens editor with the diff of those changes and lets you edit
68 it; then it applies only the diff on the tree and lets you commit
69 it or whatever. After committing, cg-unshelf or cg-shelf -u or
70 whatever will restore the state of the tree before cg-shelf was
73 * Support all the Git hooks
79 Support a GIT project inside a GIT project:
86 That means cg-update working recursively and cg-add'n'stuff
87 checking if there isn't another .git along the path of its
90 Needs more thought, especially wrt. fetching and merging
94 Probably make it a real terminal application, not just less
97 * Somehow switch between summary and full output
98 * Possibility to quickly show patches for commits
100 * Graphical commit/merge tool
102 * cg-annotate, cg-bisect
104 * Cogito's friendly patch stack handling (frontending stgit?)
105 Not sure yet if the UI improvement would be substantial
107 * More comfortable history rewriting
108 * Make it possible to do rebase-merge easily, rebasing local
109 changes against new remote head; OTOH, perhaps we just need
110 to advertise StGIT more...
111 * Easier recommitting - if I need to uncommit or change some
112 commit not at the top, let me more easily uncommit, change it
113 and recommit the rest on the top.
115 This is encouraging history rewriting which is bad as soon
116 as you publish your changes, but not before; I believe it's
117 good to offer people ways to keep their history tidy if you
118 also educate them not to do it after they publish it.
120 * Revisions numbering
121 An alternative way to identify revisions - by their sequential
122 number from the root, after merge-sorted.
127 * Rewrite in some other language
128 csh? Postscript? Befunge?
130 * Make Cogito agnostic to the underlying layer
131 Actually, I start to miss Cogito's convenient interface a lot when
132 working with other systems, and my fingers are already trained for
133 the cg-commands. But this is a very long-term goal and will take
134 a good deal of thought and work.
137 Make sure that as many projects as possible use GIT, and in
138 particular that everyone uses Cogito! ;-)
142 Stuff partially or fully involving Core GIT:
147 * Unnamed remote branches
148 Pull from an URI without cg-branch-add'ing it first.
151 * Private tags support
153 * Empty directories handling
154 Directories should be first-class index objects, not implicit.
158 The discussion about templates sharing among the toolkits somewhat
164 * Proper renaming recording
165 We need to do something about it. For now, at least properly follow
166 renames everywhere and have tree renames detection.
168 In the long term, either some Linus-envisioned smart moves detection
169 needs to be done (I'm still skeptical) or persistent file ids. One
170 trouble is keeping merge from traversing all the history.
172 * Better than three-way merging
173 Perhaps pcdv merge, let's see how will that turn out.
175 Three-way merge silently gets some cases wrong, and an army of
176 experimental VCSes developers is devising better merge algorithms
179 (pcdv merge is actually probably very similar to how the BK merge
182 * Cherrypicking support
183 At least "partial" cherrypicking support, with changing head IDs.
184 Merge up to the first dropped revision, rebase and merge on...
189 I envision each head having a revision journal - list of all preceding
190 revisions in their merge order.
192 There are two motivations. The first is possibility of fast and
193 setupless dumb server transport. But more importantly, git-rev-list
194 does not scale if you want to sort it by anything but date, and you
195 definitively want - I think the date order is confusing and e.g.
196 for cg-mkpatch simply wrong and you really need merge order instead.
197 But that requires to load all the history, then order it, and that
198 means that it takes to crunch through the whole history even when
199 showing the tip of cg-log. And that is not going to scale to
200 thousands of revisions - it already takes 2s (20s for cold cache!)
201 for the current Cogito/git repositories, which is too long for
204 So instead, every time you commit, the new revision is appended
205 to the revision journal of the given head, and when you fetch the
206 head, you bring the remote revision journal along. When you merge,
207 all the new revisions are appended between the previous head and
208 the merge head - that gives you the merge order.
210 Well, if you want to append to it, you want new revisions at the end.
211 If you want to fetch it, you want the new revisions at the start.
212 So I think the journal should look like
214 .git/revlist/master/0001
215 .git/revlist/master/0002
218 where each file contains some 1024 revisions or so, with the new
221 Needs more thought. See also
223 Jun 16 Jon Seymour [PATCH 1/1] [PROPOSAL] Add a module (repo-log.c) to log repository events.