Even more aggressive commit trimming.
[alt-git.git] / git-subtree.sh
blob03107e2251c35e0f31e4e29ab9f0fb5515079ced
1 #!/bin/bash
3 # git-subtree.sh: split/join git repositories in subdirectories of this one
5 # Copyright (C) 2009 Avery Pennarun <apenwarr@gmail.com>
7 OPTS_SPEC="\
8 git subtree split [--rejoin] [--onto rev] <commit...> -- <path>
9 git subtree merge
11 git subtree does foo and bar!
13 h,help show the help
14 q quiet
15 v verbose
16 onto= existing subtree revision to search for parent
17 rejoin merge the new branch back into HEAD
19 eval $(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@" || echo exit $?)
20 . git-sh-setup
21 require_work_tree
23 quiet=
24 command=
25 onto=
26 rejoin=
28 debug()
30 if [ -z "$quiet" ]; then
31 echo "$@" >&2
35 assert()
37 if "$@"; then
39 else
40 die "assertion failed: " "$@"
45 #echo "Options: $*"
47 while [ $# -gt 0 ]; do
48 opt="$1"
49 shift
50 case "$opt" in
51 -q) quiet=1 ;;
52 --onto) onto="$1"; shift ;;
53 --rejoin) rejoin=1 ;;
54 --) break ;;
55 esac
56 done
58 command="$1"
59 shift
60 case "$command" in
61 split|merge) ;;
62 *) die "Unknown command '$command'" ;;
63 esac
65 revs=$(git rev-parse --default HEAD --revs-only "$@") || exit $?
66 dirs="$(git rev-parse --sq --no-revs --no-flags "$@")" || exit $?
68 #echo "dirs is {$dirs}"
69 eval $(echo set -- $dirs)
70 if [ "$#" -ne 1 ]; then
71 die "Must provide exactly one subtree dir (got $#)"
73 dir="$1"
75 debug "command: {$command}"
76 debug "quiet: {$quiet}"
77 debug "revs: {$revs}"
78 debug "dir: {$dir}"
80 cache_setup()
82 cachedir="$GIT_DIR/subtree-cache/$$"
83 rm -rf "$cachedir" || die "Can't delete old cachedir: $cachedir"
84 mkdir -p "$cachedir" || die "Can't create new cachedir: $cachedir"
85 debug "Using cachedir: $cachedir" >&2
88 cache_get()
90 for oldrev in $*; do
91 if [ -r "$cachedir/$oldrev" ]; then
92 read newrev <"$cachedir/$oldrev"
93 echo $newrev
95 done
98 cache_set()
100 oldrev="$1"
101 newrev="$2"
102 if [ "$oldrev" != "latest_old" \
103 -a "$oldrev" != "latest_new" \
104 -a -e "$cachedir/$oldrev" ]; then
105 die "cache for $oldrev already exists!"
107 echo "$newrev" >"$cachedir/$oldrev"
110 find_existing_splits()
112 debug "Looking for prior splits..."
113 dir="$1"
114 revs="$2"
115 git log --grep="^git-subtree-dir: $dir\$" \
116 --pretty=format:'%s%n%n%b%nEND' "$revs" |
117 while read a b junk; do
118 case "$a" in
119 git-subtree-mainline:) main="$b" ;;
120 git-subtree-split:) sub="$b" ;;
122 if [ -n "$main" -a -n "$sub" ]; then
123 debug " Prior: $main -> $sub"
124 cache_set $main $sub
125 echo "^$main^ ^$sub^"
126 main=
127 sub=
130 esac
131 done
134 copy_commit()
136 # We're doing to set some environment vars here, so
137 # do it in a subshell to get rid of them safely later
138 git log -1 --pretty=format:'%an%n%ae%n%ad%n%cn%n%ce%n%cd%n%s%n%n%b' "$1" |
140 read GIT_AUTHOR_NAME
141 read GIT_AUTHOR_EMAIL
142 read GIT_AUTHOR_DATE
143 read GIT_COMMITTER_NAME
144 read GIT_COMMITTER_EMAIL
145 read GIT_COMMITTER_DATE
146 export GIT_AUTHOR_NAME \
147 GIT_AUTHOR_EMAIL \
148 GIT_AUTHOR_DATE \
149 GIT_COMMITTER_NAME \
150 GIT_COMMITTER_EMAIL \
151 GIT_COMMITTER_DATE
152 (echo -n '*'; cat ) | # FIXME
153 git commit-tree "$2" $3 # reads the rest of stdin
154 ) || die "Can't copy commit $1"
157 merge_msg()
159 dir="$1"
160 latest_old="$2"
161 latest_new="$3"
162 cat <<-EOF
163 Split '$dir/' into commit '$latest_new'
165 git-subtree-dir: $dir
166 git-subtree-mainline: $latest_old
167 git-subtree-split: $latest_new
171 toptree_for_commit()
173 commit="$1"
174 git log -1 --pretty=format:'%T' "$commit" -- || exit $?
177 subtree_for_commit()
179 commit="$1"
180 dir="$2"
181 git ls-tree "$commit" -- "$dir" |
182 while read mode type tree name; do
183 assert [ "$name" = "$dir" ]
184 echo $tree
185 break
186 done
189 tree_changed()
191 tree=$1
192 shift
193 if [ $# -ne 1 ]; then
194 return 0 # weird parents, consider it changed
195 else
196 ptree=$(toptree_for_commit $1)
197 if [ "$ptree" != "$tree" ]; then
198 return 0 # changed
199 else
200 return 1 # not changed
205 copy_or_skip()
207 rev="$1"
208 tree="$2"
209 newparents="$3"
210 assert [ -n "$tree" ]
212 p=""
213 for parent in $newparents; do
214 ptree=$(toptree_for_commit $parent) || exit $?
215 if [ "$ptree" = "$tree" ]; then
216 # any identical parent means this commit is unnecessary
217 echo $parent
218 return 0
219 elif [ -n "$ptree" ]; then
220 # an existing, non-identical parent is important
221 p="$p -p $parent"
223 done
225 copy_commit $rev $tree "$p" || exit $?
228 cmd_split()
230 debug "Splitting $dir..."
231 cache_setup || exit $?
233 if [ -n "$onto" ]; then
234 debug "Reading history for --onto=$onto..."
235 git rev-list $onto |
236 while read rev; do
237 # the 'onto' history is already just the subdir, so
238 # any parent we find there can be used verbatim
239 debug " cache: $rev"
240 cache_set $rev $rev
241 done
244 unrevs="$(find_existing_splits "$dir" "$revs")"
246 debug "git rev-list --reverse $revs $unrevs"
247 git rev-list --reverse --parents $revs $unrevsx |
248 while read rev parents; do
249 debug
250 debug "Processing commit: $rev"
251 exists=$(cache_get $rev)
252 if [ -n "$exists" ]; then
253 debug " prior: $exists"
254 continue
256 debug " parents: $parents"
257 newparents=$(cache_get $parents)
258 debug " newparents: $newparents"
260 tree=$(subtree_for_commit $rev "$dir")
261 debug " tree is: $tree"
262 [ -z $tree ] && continue
264 newrev=$(copy_or_skip "$rev" "$tree" "$newparents") || exit $?
265 debug " newrev is: $newrev"
266 cache_set $rev $newrev
267 cache_set latest_new $newrev
268 cache_set latest_old $rev
269 done || exit $?
270 latest_new=$(cache_get latest_new)
271 if [ -z "$latest_new" ]; then
272 die "No new revisions were found"
275 if [ -n "$rejoin" ]; then
276 debug "Merging split branch into HEAD..."
277 latest_old=$(cache_get latest_old)
278 git merge -s ours \
279 -m "$(merge_msg $dir $latest_old $latest_new)" \
280 $latest_new >&2
282 echo $latest_new
283 exit 0
286 cmd_merge()
288 die "merge command not implemented yet"
291 "cmd_$command"