4 The latest copy of this document is found at
6 http://kernel.org/git/?p=git/git.git;a=blob;hb=todo;f=TODO
12 - In 0.99.8, we will still install the backward compatible
13 symbolic links in $(bindir). These will however be removed
16 git-ssh-push and git-ssh-pull pair is not going away within
17 this timeframe, if ever. Each of these old-name commands
18 continues to invoke its old-name counterpart on the other
22 What to expect after 0.99.8
23 ===========================
25 This is written in a form of to-do list for me, so if I say
26 "accept patch", it means I do not currently plan to do that
27 myself. People interested in seeing it materialize please take
28 a hint. Also whatever I marked "Perhaps" do not have to happen
29 if ever -- only if somebody cares enough and submits a clean
36 * Help Jon Loeliger to find place in the documentation to place
39 * Accept patches from people who actually have done CVS
40 migration and update the cvs-migration documentation.
41 Link the documentation from the main git.txt page.
43 * Talk about using rsync just once at the beginning when
44 initializing a remote repository so that local packs do not
45 need to be expanded. I personally do not think we need tool
46 support for this (but see below about optimized cloning).
48 * Maybe update tutorial with a toy project that involves two or
51 * Update tutorial to cover setting up repository hooks to do
54 * Accept patches to finish missing docs.
56 * Accept patches to talk about "Whoops, it broke. What's
59 * Accept patches to make formatted tables in asciidoc to work
60 well in both html and man pages (see git-diff(1)).
66 * We might want to optimize cloning with GIT native transport
67 not to explode the pack, and store it in objects/pack instead.
68 We would need a tool to generate an idx file out of a pack
69 file for this. Also this itself may turn out to be a bad
70 idea, making the set of packs in repositories everybody has
71 different from each other.
73 * Git daemon, when deployed at kernel.org, might turn out to be
74 quite a burden, since it needs to generate customized packs
75 every time a new request comes in. It may be worthwhile to
76 precompute some packs for popular sets of heads downloaders
77 have and serve that, even if that could give more than the
78 client asks for in some cases. We will know about this soon
81 * Libification. There are many places "run once" mentality is
82 ingrained in the management of basic data structures, which
83 need to be fixed. [Matthias Urlichs is already working on
84 this: <pan.2005.10.03.20.48.52.132570@smurf.noris.de>; Post
87 * Maybe a pack optimizer.
89 Given a set of objects and a set of refs (probably a handful
90 branch heads and point release tags), find a set of packs to
91 allow reasonably minimum download for all of these classes of
92 people: (1) somebody cloning the repository from scratch, (2)
93 somebody who tends to follow the master branch head reasonably
94 closely, (3) somebody who tends to follow only the point
97 * Maybe an Emacs VC backend.
99 * 'git split-projects'? This requires updated 'git-rev-list' to
100 skip irrelevant commits.
101 Message-ID: <Pine.LNX.4.63.0509221617300.23242@iabervon.org>
103 * Look at libified GNU diff CVS seems to use, or libxdiff.
104 [Daniel has his own diff tool almost ready to start
105 integrating and testing; Post 1.0]
107 * Accept patches to fetch multiple objects by HTTP in parallel.
108 [Preferably before 1.0]
110 * Plug-in file-level merges [Post 1.0].
116 * Review the Makefile variables and exporting rules for them,
117 while looking at prefix passing by Kai Ruemmler.
119 * Review the 'sparce object database' change by Linus and move
120 the first phase of it to the "master" branch.
122 * Decide on mmap(). I am inclined to just stick to mmap
123 replacement by Johannes Schindelin and do nothing else right
126 * Revisit Santi's patch to move commit temorary files out of the
127 working tree toplevel. More generally, review the use of
128 temporary files again. Assuming $GIT_DIR is writable is
129 more acceptable, but the working tree toplevel may not be in
130 a rare usage pattern.
132 * Perhaps accept more "want"s in upload-pack and do something
133 intelligent about it.
135 * Perhaps detect cloning request in upload-pack and cache the
136 result for next cloning request until any of our refs change.
138 * Perhaps send less "want"s from fetch-pack.
140 * Encourage concrete proposals to commit log message templates
141 we discussed some time ago.
143 * Accept patches to cause "read-tree -u" delete a directory when
146 * Perhaps accept patches to do undo/redo.
148 * Perhaps accept patch to optionally allow '--fuzz' in
151 * Allow 'git apply' to accept GNU diff 2.7 output that forgets
152 to say '\No newline' if both input ends with incomplete
155 * What to do with TABs and LFs in pathnames without breaking GNU
158 * Maybe grok PGP signed text/plain in applymbox as well.
160 * Perhaps a tool to revert a single file to pre-modification
161 state? People with BK background know this operation as
162 'clean'. 'git checkout [-f] ent [path...]' was suggested by
163 Matthias Urlichs which sounds a natural extention to what the
164 command currently does.
166 * Enhance "git repack" to not always use --all; this would be
167 handy if the repository contains wagging heads like "pu" in
170 * Internally split the project into non-doc and doc parts; add
171 an extra root for the doc part and merge from it; move the
172 internal doc source to a separate repository, like the +Meta
173 repository; experiment if this results in a reasonable
174 workflow, and document it in howto form if it does.
176 The point is to make it possible to fork that part off to
177 somebody else; then I do not have to maintain Documentation
178 directory myself anymore, just like I simply slurp the latest
179 gitk from Paul and not worry about it ;-).
181 * Make rebase restartable; instead of skipping what cannot be
182 automatically forward ported, leave the conflicts in the work
183 tree, have the user resolve it, and then restart from where it
186 * Output full path in the "git-rev-list --objects" output, not
187 just the basename, and see the improved clustering results in
188 better packing [Tried, but did not work out well].
190 * Updated git-changes-script Jeff Garzik needs [Inquiry for
191 external spec sent out with a quick hack. Will know if that
192 is what he needs soon enough].
198 * short SHA1 naming is not enforcing uniqueness. Should fix [DONE].
200 * 'git repack' can be DOSed. Should fix [DONE].
202 * Stop installing the old-name symlinks [POSTPONED, but before 1.0].
204 * 'git merge-projects'?
206 Subject: Re: Merges without bases
207 References: <1125004228.4110.20.camel@localhost.localdomain>
208 Date: Thu, 25 Aug 2005 15:26:36 -0700
209 Message-ID: <7vvf1tps9v.fsf@assigned-by-dhcp.cox.net>
211 * 'git lost-and-found'? Link dangling commits found by
212 fsck-objects under $GIT_DIR/refs/lost-found/. Then
213 show-branch or gitk can be used to find any lost commit. [A
214 feeler patch sent out. Very underwhelming response X-<.]
216 Do not name it /lost+found/; that would probably confuse
217 things that mistake it a mount point (not our code but
220 * Add simple globbing rules to git-show-branch so that I can
221 say 'git show-branch --heads "ko-*"' (ko-master, ko-pu, and
222 ko-rc are in refs/tags/).
224 * We would want test scripts for the relative directory path
225 stuff Linus has been working on. So far, the following
226 commands should be usable with relative directory paths:
236 * In a freashly created empty repository, `git fetch foo:bar`
237 works OK, but `git checkout bar` afterwards does not (missing
projects / git / blob