The third batch
[git.git] / git-filter-branch.sh
blob5c5afa2b985e2e524601518bbb8fb8d3c8efda61
1 #!/bin/sh
3 # Rewrite revision history
4 # Copyright (c) Petr Baudis, 2006
5 # Minimal changes to "port" it to core-git (c) Johannes Schindelin, 2007
7 # Lets you rewrite the revision history of the current branch, creating
8 # a new branch. You can specify a number of filters to modify the commits,
9 # files and trees.
11 # The following functions will also be available in the commit filter:
13 functions=$(cat << \EOF
14 EMPTY_TREE=$(git hash-object -t tree /dev/null)
16 warn () {
17 echo "$*" >&2
20 map()
22 # if it was not rewritten, take the original
23 if test -r "$workdir/../map/$1"
24 then
25 cat "$workdir/../map/$1"
26 else
27 echo "$1"
31 # if you run 'skip_commit "$@"' in a commit filter, it will print
32 # the (mapped) parents, effectively skipping the commit.
34 skip_commit()
36 shift;
37 while [ -n "$1" ];
39 shift;
40 map "$1";
41 shift;
42 done;
45 # if you run 'git_commit_non_empty_tree "$@"' in a commit filter,
46 # it will skip commits that leave the tree untouched, commit the other.
47 git_commit_non_empty_tree()
49 if test $# = 3 && test "$1" = $(git rev-parse "$3^{tree}"); then
50 map "$3"
51 elif test $# = 1 && test "$1" = $EMPTY_TREE; then
53 else
54 git commit-tree "$@"
57 # override die(): this version puts in an extra line break, so that
58 # the progress is still visible
60 die()
62 echo >&2
63 echo "$*" >&2
64 exit 1
66 EOF
69 eval "$functions"
71 finish_ident() {
72 # Ensure non-empty id name.
73 echo "case \"\$GIT_$1_NAME\" in \"\") GIT_$1_NAME=\"\${GIT_$1_EMAIL%%@*}\" && export GIT_$1_NAME;; esac"
74 # And make sure everything is exported.
75 echo "export GIT_$1_NAME"
76 echo "export GIT_$1_EMAIL"
77 echo "export GIT_$1_DATE"
80 set_ident () {
81 parse_ident_from_commit author AUTHOR committer COMMITTER
82 finish_ident AUTHOR
83 finish_ident COMMITTER
86 USAGE="[--setup <command>] [--subdirectory-filter <directory>] [--env-filter <command>]
87 [--tree-filter <command>] [--index-filter <command>]
88 [--parent-filter <command>] [--msg-filter <command>]
89 [--commit-filter <command>] [--tag-name-filter <command>]
90 [--original <namespace>]
91 [-d <directory>] [-f | --force] [--state-branch <branch>]
92 [--] [<rev-list options>...]"
94 OPTIONS_SPEC=
95 . git-sh-setup
97 if [ "$(is_bare_repository)" = false ]; then
98 require_clean_work_tree 'rewrite branches'
101 tempdir=.git-rewrite
102 filter_setup=
103 filter_env=
104 filter_tree=
105 filter_index=
106 filter_parent=
107 filter_msg=cat
108 filter_commit=
109 filter_tag_name=
110 filter_subdir=
111 state_branch=
112 orig_namespace=refs/original/
113 force=
114 prune_empty=
115 remap_to_ancestor=
116 while :
118 case "$1" in
120 shift
121 break
123 --force|-f)
124 shift
125 force=t
126 continue
128 --remap-to-ancestor)
129 # deprecated ($remap_to_ancestor is set now automatically)
130 shift
131 remap_to_ancestor=t
132 continue
134 --prune-empty)
135 shift
136 prune_empty=t
137 continue
142 break;
143 esac
145 # all switches take one argument
146 ARG="$1"
147 case "$#" in 1) usage ;; esac
148 shift
149 OPTARG="$1"
150 shift
152 case "$ARG" in
154 tempdir="$OPTARG"
156 --setup)
157 filter_setup="$OPTARG"
159 --subdirectory-filter)
160 filter_subdir="$OPTARG"
161 remap_to_ancestor=t
163 --env-filter)
164 filter_env="$OPTARG"
166 --tree-filter)
167 filter_tree="$OPTARG"
169 --index-filter)
170 filter_index="$OPTARG"
172 --parent-filter)
173 filter_parent="$OPTARG"
175 --msg-filter)
176 filter_msg="$OPTARG"
178 --commit-filter)
179 filter_commit="$functions; $OPTARG"
181 --tag-name-filter)
182 filter_tag_name="$OPTARG"
184 --original)
185 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
187 --state-branch)
188 state_branch="$OPTARG"
191 usage
193 esac
194 done
196 case "$prune_empty,$filter_commit" in
198 filter_commit='git commit-tree "$@"';;
200 filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
204 die "Cannot set --prune-empty and --commit-filter at the same time"
205 esac
207 case "$force" in
209 rm -rf "$tempdir"
212 test -d "$tempdir" &&
213 die "$tempdir already exists, please remove it"
214 esac
215 orig_dir=$(pwd)
216 mkdir -p "$tempdir/t" &&
217 tempdir="$(cd "$tempdir"; pwd)" &&
218 cd "$tempdir/t" &&
219 workdir="$(pwd)" ||
220 die ""
222 # Remove tempdir on exit
223 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
225 ORIG_GIT_DIR="$GIT_DIR"
226 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
227 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
228 ORIG_GIT_AUTHOR_NAME="$GIT_AUTHOR_NAME"
229 ORIG_GIT_AUTHOR_EMAIL="$GIT_AUTHOR_EMAIL"
230 ORIG_GIT_AUTHOR_DATE="$GIT_AUTHOR_DATE"
231 ORIG_GIT_COMMITTER_NAME="$GIT_COMMITTER_NAME"
232 ORIG_GIT_COMMITTER_EMAIL="$GIT_COMMITTER_EMAIL"
233 ORIG_GIT_COMMITTER_DATE="$GIT_COMMITTER_DATE"
235 GIT_WORK_TREE=.
236 export GIT_DIR GIT_WORK_TREE
238 # Make sure refs/original is empty
239 git for-each-ref > "$tempdir"/backup-refs || exit
240 while read sha1 type name
242 case "$force,$name" in
243 ,$orig_namespace*)
244 die "Cannot create a new backup.
245 A previous backup already exists in $orig_namespace
246 Force overwriting the backup with -f"
248 t,$orig_namespace*)
249 git update-ref -d "$name" $sha1
251 esac
252 done < "$tempdir"/backup-refs
254 # The refs should be updated if their heads were rewritten
255 git rev-parse --no-flags --revs-only --symbolic-full-name \
256 --default HEAD "$@" > "$tempdir"/raw-refs || exit
257 while read ref
259 case "$ref" in ^?*) continue ;; esac
261 if git rev-parse --verify "$ref"^0 >/dev/null 2>&1
262 then
263 echo "$ref"
264 else
265 warn "WARNING: not rewriting '$ref' (not a committish)"
267 done >"$tempdir"/heads <"$tempdir"/raw-refs
269 test -s "$tempdir"/heads ||
270 die "You must specify a ref to rewrite."
272 GIT_INDEX_FILE="$(pwd)/../index"
273 export GIT_INDEX_FILE
275 # map old->new commit ids for rewriting parents
276 mkdir ../map || die "Could not create map/ directory"
278 if test -n "$state_branch"
279 then
280 state_commit=$(git rev-parse --no-flags --revs-only "$state_branch")
281 if test -n "$state_commit"
282 then
283 echo "Populating map from $state_branch ($state_commit)" 1>&2
284 perl -e'open(MAP, "-|", "git show $ARGV[0]:filter.map") or die;
285 while (<MAP>) {
286 m/(.*):(.*)/ or die;
287 open F, ">../map/$1" or die;
288 print F "$2" or die;
289 close(F) or die;
291 close(MAP) or die;' "$state_commit" \
292 || die "Unable to load state from $state_branch:filter.map"
293 else
294 echo "Branch $state_branch does not exist. Will create" 1>&2
298 # we need "--" only if there are no path arguments in $@
299 nonrevs=$(git rev-parse --no-revs "$@") || exit
300 if test -z "$nonrevs"
301 then
302 dashdash=--
303 else
304 dashdash=
305 remap_to_ancestor=t
308 git rev-parse --revs-only "$@" >../parse
310 case "$filter_subdir" in
312 eval set -- "$(git rev-parse --sq --no-revs "$@")"
315 eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
316 "$filter_subdir")"
318 esac
320 git rev-list --reverse --topo-order --default HEAD \
321 --parents --simplify-merges --stdin "$@" <../parse >../revs ||
322 die "Could not get the commits"
323 commits=$(wc -l <../revs | tr -d " ")
325 test $commits -eq 0 && die_with_status 2 "Found nothing to rewrite"
327 # Rewrite the commits
328 report_progress ()
330 if test -n "$progress" &&
331 test $git_filter_branch__commit_count -gt $next_sample_at
332 then
333 count=$git_filter_branch__commit_count
335 now=$(date +%s)
336 elapsed=$(($now - $start_timestamp))
337 remaining=$(( ($commits - $count) * $elapsed / $count ))
338 if test $elapsed -gt 0
339 then
340 next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
341 else
342 next_sample_at=$(($next_sample_at + 1))
344 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
346 printf "\rRewrite $commit ($count/$commits)$progress "
349 git_filter_branch__commit_count=0
351 progress= start_timestamp=
352 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
353 then
354 next_sample_at=0
355 progress="dummy to ensure this is not empty"
356 start_timestamp=$(date '+%s')
359 if test -n "$filter_index" ||
360 test -n "$filter_tree" ||
361 test -n "$filter_subdir"
362 then
363 need_index=t
364 else
365 need_index=
368 eval "$filter_setup" < /dev/null ||
369 die "filter setup failed: $filter_setup"
371 while read commit parents; do
372 git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
374 report_progress
375 test -f "$workdir"/../map/$commit && continue
377 case "$filter_subdir" in
379 if test -n "$need_index"
380 then
381 GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
385 # The commit may not have the subdirectory at all
386 err=$(GIT_ALLOW_NULL_SHA1=1 \
387 git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
388 if ! git rev-parse -q --verify $commit:"$filter_subdir"
389 then
390 rm -f "$GIT_INDEX_FILE"
391 else
392 echo >&2 "$err"
393 false
396 esac || die "Could not initialize the index"
398 GIT_COMMIT=$commit
399 export GIT_COMMIT
400 git cat-file commit "$commit" >../commit ||
401 die "Cannot read commit $commit"
403 eval "$(set_ident <../commit)" ||
404 die "setting author/committer failed for commit $commit"
405 eval "$filter_env" < /dev/null ||
406 die "env filter failed: $filter_env"
408 if [ "$filter_tree" ]; then
409 git checkout-index -f -u -a ||
410 die "Could not checkout the index"
411 # files that $commit removed are now still in the working tree;
412 # remove them, else they would be added again
413 git clean -d -q -f -x
414 eval "$filter_tree" < /dev/null ||
415 die "tree filter failed: $filter_tree"
418 git diff-index -r --name-only --ignore-submodules $commit -- &&
419 git ls-files --others
420 ) > "$tempdir"/tree-state || exit
421 git update-index --add --replace --remove --stdin \
422 < "$tempdir"/tree-state || exit
425 eval "$filter_index" < /dev/null ||
426 die "index filter failed: $filter_index"
428 parentstr=
429 for parent in $parents; do
430 for reparent in $(map "$parent"); do
431 case "$parentstr " in
432 *" -p $reparent "*)
435 parentstr="$parentstr -p $reparent"
437 esac
438 done
439 done
440 if [ "$filter_parent" ]; then
441 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
442 die "parent filter failed: $filter_parent"
446 while IFS='' read -r header_line && test -n "$header_line"
448 # skip header lines...
450 done
451 # and output the actual commit message
453 } <../commit |
454 eval "$filter_msg" > ../message ||
455 die "msg filter failed: $filter_msg"
457 if test -n "$need_index"
458 then
459 tree=$(git write-tree)
460 else
461 tree=$(git rev-parse "$commit^{tree}")
463 workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
464 "$tree" $parentstr < ../message > ../map/$commit ||
465 die "could not write rewritten commit"
466 done <../revs
468 # If we are filtering for paths, as in the case of a subdirectory
469 # filter, it is possible that a specified head is not in the set of
470 # rewritten commits, because it was pruned by the revision walker.
471 # Ancestor remapping fixes this by mapping these heads to the unique
472 # nearest ancestor that survived the pruning.
474 if test "$remap_to_ancestor" = t
475 then
476 while read ref
478 sha1=$(git rev-parse "$ref"^0)
479 test -f "$workdir"/../map/$sha1 && continue
480 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
481 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
482 done < "$tempdir"/heads
485 # Finally update the refs
487 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
488 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
489 echo
490 while read ref
492 # avoid rewriting a ref twice
493 test -f "$orig_namespace$ref" && continue
495 sha1=$(git rev-parse "$ref"^0)
496 rewritten=$(map $sha1)
498 test $sha1 = "$rewritten" &&
499 warn "WARNING: Ref '$ref' is unchanged" &&
500 continue
502 case "$rewritten" in
504 echo "Ref '$ref' was deleted"
505 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
506 die "Could not delete $ref"
508 $_x40)
509 echo "Ref '$ref' was rewritten"
510 if ! git update-ref -m "filter-branch: rewrite" \
511 "$ref" $rewritten $sha1 2>/dev/null; then
512 if test $(git cat-file -t "$ref") = tag; then
513 if test -z "$filter_tag_name"; then
514 warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
515 warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
517 else
518 die "Could not rewrite $ref"
523 # NEEDSWORK: possibly add -Werror, making this an error
524 warn "WARNING: '$ref' was rewritten into multiple commits:"
525 warn "$rewritten"
526 warn "WARNING: Ref '$ref' points to the first one now."
527 rewritten=$(echo "$rewritten" | head -n 1)
528 git update-ref -m "filter-branch: rewrite to first" \
529 "$ref" $rewritten $sha1 ||
530 die "Could not rewrite $ref"
532 esac
533 git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
534 exit
535 done < "$tempdir"/heads
537 # TODO: This should possibly go, with the semantics that all positive given
538 # refs are updated, and their original heads stored in refs/original/
539 # Filter tags
541 if [ "$filter_tag_name" ]; then
542 git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
543 while read sha1 type ref; do
544 ref="${ref#refs/tags/}"
545 # XXX: Rewrite tagged trees as well?
546 if [ "$type" != "commit" -a "$type" != "tag" ]; then
547 continue;
550 if [ "$type" = "tag" ]; then
551 # Dereference to a commit
552 sha1t="$sha1"
553 sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
556 [ -f "../map/$sha1" ] || continue
557 new_sha1="$(cat "../map/$sha1")"
558 GIT_COMMIT="$sha1"
559 export GIT_COMMIT
560 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
561 die "tag name filter failed: $filter_tag_name"
563 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
565 if [ "$type" = "tag" ]; then
566 new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
567 "$new_sha1" "$new_ref"
568 git cat-file tag "$ref" |
569 sed -n \
570 -e '1,/^$/{
571 /^object /d
572 /^type /d
573 /^tag /d
574 }' \
575 -e '/^-----BEGIN PGP SIGNATURE-----/q' \
576 -e 'p' ) |
577 git hash-object -t tag -w --stdin) ||
578 die "Could not create new tag object for $ref"
579 if git cat-file tag "$ref" | \
580 sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
581 then
582 warn "gpg signature stripped from tag object $sha1t"
586 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
587 die "Could not write tag $new_ref"
588 done
591 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
592 unset GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL GIT_AUTHOR_DATE
593 unset GIT_COMMITTER_NAME GIT_COMMITTER_EMAIL GIT_COMMITTER_DATE
594 test -z "$ORIG_GIT_DIR" || {
595 GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
597 test -z "$ORIG_GIT_WORK_TREE" || {
598 GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
599 export GIT_WORK_TREE
601 test -z "$ORIG_GIT_INDEX_FILE" || {
602 GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
603 export GIT_INDEX_FILE
605 test -z "$ORIG_GIT_AUTHOR_NAME" || {
606 GIT_AUTHOR_NAME="$ORIG_GIT_AUTHOR_NAME" &&
607 export GIT_AUTHOR_NAME
609 test -z "$ORIG_GIT_AUTHOR_EMAIL" || {
610 GIT_AUTHOR_EMAIL="$ORIG_GIT_AUTHOR_EMAIL" &&
611 export GIT_AUTHOR_EMAIL
613 test -z "$ORIG_GIT_AUTHOR_DATE" || {
614 GIT_AUTHOR_DATE="$ORIG_GIT_AUTHOR_DATE" &&
615 export GIT_AUTHOR_DATE
617 test -z "$ORIG_GIT_COMMITTER_NAME" || {
618 GIT_COMMITTER_NAME="$ORIG_GIT_COMMITTER_NAME" &&
619 export GIT_COMMITTER_NAME
621 test -z "$ORIG_GIT_COMMITTER_EMAIL" || {
622 GIT_COMMITTER_EMAIL="$ORIG_GIT_COMMITTER_EMAIL" &&
623 export GIT_COMMITTER_EMAIL
625 test -z "$ORIG_GIT_COMMITTER_DATE" || {
626 GIT_COMMITTER_DATE="$ORIG_GIT_COMMITTER_DATE" &&
627 export GIT_COMMITTER_DATE
630 if test -n "$state_branch"
631 then
632 echo "Saving rewrite state to $state_branch" 1>&2
633 state_blob=$(
634 perl -e'opendir D, "../map" or die;
635 open H, "|-", "git hash-object -w --stdin" or die;
636 foreach (sort readdir(D)) {
637 next if m/^\.\.?$/;
638 open F, "<../map/$_" or die;
639 chomp($f = <F>);
640 print H "$_:$f\n" or die;
642 close(H) or die;' || die "Unable to save state")
643 state_tree=$(printf '100644 blob %s\tfilter.map\n' "$state_blob" | git mktree)
644 if test -n "$state_commit"
645 then
646 state_commit=$(echo "Sync" | git commit-tree "$state_tree" -p "$state_commit")
647 else
648 state_commit=$(echo "Sync" | git commit-tree "$state_tree" )
650 git update-ref "$state_branch" "$state_commit"
653 cd "$orig_dir"
654 rm -rf "$tempdir"
656 trap - 0
658 if [ "$(is_bare_repository)" = false ]; then
659 git read-tree -u -m HEAD || exit
662 exit 0