1 What to expect after 0.99.6
2 ===========================
4 This is written in a form of to-do list for me, so if I say
5 "accept patch", it means I do not currently plan to do that
6 myself. People interested in seeing it materialize please take
7 a hint. The latest copy of this document is found at
9 http://kernel.org/git/?p=git/git.git;a=blob;hb=todo;f=TODO
14 - All non-binary commands will lose -script suffix in
15 $(bindir). The source to git-foo will be either git-foo.sh
16 or git-foo.perl in the source tree, and the documentation
17 will be in Documentation/git-foo.txt.
19 - The commands whose names have 'cache' to mean 'index file'
20 will get 'cache' in their names replaced with 'index'. For
21 git-fsck-cache and git-convert-cache, 'cache' will be
22 replaced with 'objects'.
24 - The commit walkers will have 'pull' in their names replaced
25 with 'fetch'. 'git-ssh-push' will become 'git-ssh-upload'.
27 - We continue to follow the convention to name the C source
28 file that contains the main program of 'git-foo' command
29 'foo.c'. That means we will have 'fsck-objects.c', for
32 - At this moment, I am not planning to rename the symbols used
33 in programs, nor any library sources. "cache.h" will stay
34 "cache.h", so does "read-cache.c". "struct cache_entry" and
35 "ce_match_stat()" will keep their names. We _might_ want to
36 rename them in later rounds but not right now.
38 - In 0.99.7, all renamed commands will have symbolic links in
39 $(bindir) so that old names continue to work. These backward
40 compatible symlinks will not be present in documentation,
41 though. Especially, the main documentation, git(7) will talk
42 about the new names. Old environment names defined in
43 gitenv() will also be removed in this release.
45 Tentatively we aim to do this on Sep 17th.
47 - In 0.99.8, we do not install these backward compatible
48 symbolic links in $(bindir) anymore. The Makefile will have
49 a target to remove old symlinks from $(DESTDIR)$(bindir) you
50 can run manually to help you clean things up.
52 The timeframe for this is around Oct 1st, but I could be
53 talked into delaying the symlink removal if Porcelain people
54 find this schedule too tight.
60 * Accept patches from people who actually have done CVS
61 migration and update the cvs-migration documentation.
62 Link the documentation from the main git.txt page.
64 * Accept patches from people who were hit by shiny blue bat to
65 update the SubmittingPatches.
67 * Talk about using rsync just once at the beginning when
68 initializing a remote repository so that local packs do not
69 need to be expanded. I personally do not think we need tool
70 support for this (but see below about optimized cloning).
72 * Maybe update tutorial with a toy project that involves two or
75 * Update tutorial to cover setting up repository hooks to do
78 * Accept patches to finish missing docs.
84 * Tony Luck reported an unfortunate glitch in the 3-way merge.
85 Encourage discussions to come up with a not-so-expensive way
86 to catch the kind of ambiguities that led to his misery.
87 [Deathmatch between Daniel's and Fredrik's ongoing.]
89 * HPA has two projects, klibc and klibc-kbuild, that have large
90 set of overlapping files in different paths (i.e. one has many
91 renames from the other). There currently is no way for git to
92 help keep these two trees in sync, merging criss-cross between
93 them. The merge logic should be able to take advantage of
94 rename/copy detection smarts git-diff-* family has. Linus,
95 me, and Daniel outlined a smarter merge strategy for this.
98 * To make it easier to experiment with different merge
99 strategies, make git-merge driver that will run merge backends
100 for the best merge [DONE].
102 * We might want to optimize cloning with GIT native transport
103 not to explode the pack, and store it in objects/pack instead.
104 We would need a tool to generate an idx file out of a pack
105 file for this. Also this itself may turn out to be a bad
106 idea, making the set of packs in repositories everybody has
107 different from each other.
109 * Maybe a pack optimizer.
111 * Maybe an Emacs VC backend.
117 * Define semantics for relative pathnames in objects/info/alternates.
118 Either relative to 'objects', or relative to the top of project tree
119 ;the latter is consistent with GIT_ALTERNATE_OBJECT_DIRECTORIES.
120 Perhaps would end up doing the former because we would want to be
121 consistent between a naked repository and a repository with working
122 tree [RFC-PATCH issued].
124 * Tool renames. Give fixup patch for at least Cogito and StGIT
127 * Have Daniel's read-tree graduate from "pu" after plugging
130 * Implement a merge backend using Daniel's read-tree [DONE].
132 * Rename 'git-merge-resolve' to 'git-merge-stupid' and
133 'git-merge-multibase' to 'git-merge-resolve'; the former is
134 for historical curiosity and comparison purposes only and not
135 to be used in real applications so remove it from the default
136 strategies list [DONE].
138 * Accept Fredrik merge after renaming it (I want to name the
139 driver 'git merge'). Suggest where to place *.py stuff --
140 probably in $(share)/git-core/ and add Makefile entry for
143 * Encourage concrete proposals to commit log message templates
144 we discussed some time ago.
146 * Bug Martin for archimport script documentation [DONE].
148 * More portability [IN PROGRESS].
150 * Accept patches to cause "read-tree -u" delete a directory when
153 * Perhaps accept patches to introduce the concept of "patch flow
154 expressed as ref mappings" Josef has been advocating about.
156 * Perhaps accept patches to do undo/redo.
158 * Perhaps accept patch to optionally allow '--fuzz' in
161 * Maybe grok PGP signed text/plain in applymbox as well.
163 * Perhaps a tool to revert a single file to pre-modification
164 state? git-cat-file blob `git-ls-files | grep foo` >foo or
165 git-cat-file blob `git-ls-tree HEAD foo` >foo? What should
166 the command be called? git-revert is taken so is
169 * A tool to detect, show and prune already merged topic
172 * "Lost and Found"; link dangling commits found by fsck-objects
173 under $GIT_DIR/refs/lost-found/. Then show-branch or gitk can
174 be used to find any lost commit.
176 Do not name it /lost+found/; that would probably confuse
177 things that mistake it a mount point (not our code but
180 * Enhance "git repack" to not always use --all; this would be
181 handy if the repository contains wagging heads like "pu" in
184 * Internally split the project into non-doc and doc parts; add
185 an extra root for the doc part and merge from it; move the
186 internal doc source to a separate repository, like the +Meta
187 repository; experiment if this results in a reasonable
188 workflow, and document it in howto form if it does.
190 * Option to limit rename detection for more than N paths.
192 * Marking copy seems to be broken when A,B => B,C rename
193 happens; we incorrectly say B stays, which makes the new B
194 copy of old A, not rename [DONE].
196 * Make rebase restartable; instead of skipping what cannot be
197 automatically forward ported, leave the conflicts in the work
198 tree, have the user resolve it, and then restart from where it
201 * Output full path in the "git-rev-list --objects" output, not
202 just the basename, and see the improved clustering results in
203 better packing [Tried, but did not work out well].
209 * Deal with CDPATH [DONE].
211 * Teach 'git format-patch' about 'git cherry' [DONE].
213 * Make 'git bisect' easier to use by logging the earlier
214 good/bad choices and make it replayable [DONE].
216 * Perhaps "git branch -d" to delete a branch. We probably need
217 some way to say "delete this branch because everything in this
218 should be merged in the master branch" and refuse the removal
221 * We would want test scripts for the relative directory path
222 stuff Linus has been working on. So far, the following
223 commands should be usable with relative directory paths:
233 * In a freashly created empty repository, `git fetch foo:bar`
234 works OK, but `git checkout bar` afterwards does not (missing