clone: add --sparse mode
[git.git] / git-filter-branch.sh
blobfea79646172184c630508908dd6eb61d3470074d
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_$1_DATE"
80 set_ident () {
81 parse_ident_from_commit author AUTHOR committer COMMITTER
82 finish_ident AUTHOR
83 finish_ident COMMITTER
86 if test -z "$FILTER_BRANCH_SQUELCH_WARNING$GIT_TEST_DISALLOW_ABBREVIATED_OPTIONS"
87 then
88 cat <<EOF
89 WARNING: git-filter-branch has a glut of gotchas generating mangled history
90 rewrites. Hit Ctrl-C before proceeding to abort, then use an
91 alternative filtering tool such as 'git filter-repo'
92 (https://github.com/newren/git-filter-repo/) instead. See the
93 filter-branch manual page for more details; to squelch this warning,
94 set FILTER_BRANCH_SQUELCH_WARNING=1.
95 EOF
96 sleep 10
97 printf "Proceeding with filter-branch...\n\n"
100 USAGE="[--setup <command>] [--subdirectory-filter <directory>] [--env-filter <command>]
101 [--tree-filter <command>] [--index-filter <command>]
102 [--parent-filter <command>] [--msg-filter <command>]
103 [--commit-filter <command>] [--tag-name-filter <command>]
104 [--original <namespace>]
105 [-d <directory>] [-f | --force] [--state-branch <branch>]
106 [--] [<rev-list options>...]"
108 OPTIONS_SPEC=
109 . git-sh-setup
111 if [ "$(is_bare_repository)" = false ]; then
112 require_clean_work_tree 'rewrite branches'
115 tempdir=.git-rewrite
116 filter_setup=
117 filter_env=
118 filter_tree=
119 filter_index=
120 filter_parent=
121 filter_msg=cat
122 filter_commit=
123 filter_tag_name=
124 filter_subdir=
125 state_branch=
126 orig_namespace=refs/original/
127 force=
128 prune_empty=
129 remap_to_ancestor=
130 while :
132 case "$1" in
134 shift
135 break
137 --force|-f)
138 shift
139 force=t
140 continue
142 --remap-to-ancestor)
143 # deprecated ($remap_to_ancestor is set now automatically)
144 shift
145 remap_to_ancestor=t
146 continue
148 --prune-empty)
149 shift
150 prune_empty=t
151 continue
156 break;
157 esac
159 # all switches take one argument
160 ARG="$1"
161 case "$#" in 1) usage ;; esac
162 shift
163 OPTARG="$1"
164 shift
166 case "$ARG" in
168 tempdir="$OPTARG"
170 --setup)
171 filter_setup="$OPTARG"
173 --subdirectory-filter)
174 filter_subdir="$OPTARG"
175 remap_to_ancestor=t
177 --env-filter)
178 filter_env="$OPTARG"
180 --tree-filter)
181 filter_tree="$OPTARG"
183 --index-filter)
184 filter_index="$OPTARG"
186 --parent-filter)
187 filter_parent="$OPTARG"
189 --msg-filter)
190 filter_msg="$OPTARG"
192 --commit-filter)
193 filter_commit="$functions; $OPTARG"
195 --tag-name-filter)
196 filter_tag_name="$OPTARG"
198 --original)
199 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
201 --state-branch)
202 state_branch="$OPTARG"
205 usage
207 esac
208 done
210 case "$prune_empty,$filter_commit" in
212 filter_commit='git commit-tree "$@"';;
214 filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
218 die "Cannot set --prune-empty and --commit-filter at the same time"
219 esac
221 case "$force" in
223 rm -rf "$tempdir"
226 test -d "$tempdir" &&
227 die "$tempdir already exists, please remove it"
228 esac
229 orig_dir=$(pwd)
230 mkdir -p "$tempdir/t" &&
231 tempdir="$(cd "$tempdir"; pwd)" &&
232 cd "$tempdir/t" &&
233 workdir="$(pwd)" ||
234 die ""
236 # Remove tempdir on exit
237 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
239 ORIG_GIT_DIR="$GIT_DIR"
240 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
241 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
242 ORIG_GIT_AUTHOR_NAME="$GIT_AUTHOR_NAME"
243 ORIG_GIT_AUTHOR_EMAIL="$GIT_AUTHOR_EMAIL"
244 ORIG_GIT_AUTHOR_DATE="$GIT_AUTHOR_DATE"
245 ORIG_GIT_COMMITTER_NAME="$GIT_COMMITTER_NAME"
246 ORIG_GIT_COMMITTER_EMAIL="$GIT_COMMITTER_EMAIL"
247 ORIG_GIT_COMMITTER_DATE="$GIT_COMMITTER_DATE"
249 GIT_WORK_TREE=.
250 export GIT_DIR GIT_WORK_TREE
252 # Make sure refs/original is empty
253 git for-each-ref > "$tempdir"/backup-refs || exit
254 while read sha1 type name
256 case "$force,$name" in
257 ,$orig_namespace*)
258 die "Cannot create a new backup.
259 A previous backup already exists in $orig_namespace
260 Force overwriting the backup with -f"
262 t,$orig_namespace*)
263 git update-ref -d "$name" $sha1
265 esac
266 done < "$tempdir"/backup-refs
268 # The refs should be updated if their heads were rewritten
269 git rev-parse --no-flags --revs-only --symbolic-full-name \
270 --default HEAD "$@" > "$tempdir"/raw-refs || exit
271 while read ref
273 case "$ref" in ^?*) continue ;; esac
275 if git rev-parse --verify "$ref"^0 >/dev/null 2>&1
276 then
277 echo "$ref"
278 else
279 warn "WARNING: not rewriting '$ref' (not a committish)"
281 done >"$tempdir"/heads <"$tempdir"/raw-refs
283 test -s "$tempdir"/heads ||
284 die "You must specify a ref to rewrite."
286 GIT_INDEX_FILE="$(pwd)/../index"
287 export GIT_INDEX_FILE
289 # map old->new commit ids for rewriting parents
290 mkdir ../map || die "Could not create map/ directory"
292 if test -n "$state_branch"
293 then
294 state_commit=$(git rev-parse --no-flags --revs-only "$state_branch")
295 if test -n "$state_commit"
296 then
297 echo "Populating map from $state_branch ($state_commit)" 1>&2
298 perl -e'open(MAP, "-|", "git show $ARGV[0]:filter.map") or die;
299 while (<MAP>) {
300 m/(.*):(.*)/ or die;
301 open F, ">../map/$1" or die;
302 print F "$2" or die;
303 close(F) or die;
305 close(MAP) or die;' "$state_commit" \
306 || die "Unable to load state from $state_branch:filter.map"
307 else
308 echo "Branch $state_branch does not exist. Will create" 1>&2
312 # we need "--" only if there are no path arguments in $@
313 nonrevs=$(git rev-parse --no-revs "$@") || exit
314 if test -z "$nonrevs"
315 then
316 dashdash=--
317 else
318 dashdash=
319 remap_to_ancestor=t
322 git rev-parse --revs-only "$@" >../parse
324 case "$filter_subdir" in
326 eval set -- "$(git rev-parse --sq --no-revs "$@")"
329 eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
330 "$filter_subdir")"
332 esac
334 git rev-list --reverse --topo-order --default HEAD \
335 --parents --simplify-merges --stdin "$@" <../parse >../revs ||
336 die "Could not get the commits"
337 commits=$(wc -l <../revs | tr -d " ")
339 test $commits -eq 0 && die_with_status 2 "Found nothing to rewrite"
341 # Rewrite the commits
342 report_progress ()
344 if test -n "$progress" &&
345 test $git_filter_branch__commit_count -gt $next_sample_at
346 then
347 count=$git_filter_branch__commit_count
349 now=$(date +%s)
350 elapsed=$(($now - $start_timestamp))
351 remaining=$(( ($commits - $count) * $elapsed / $count ))
352 if test $elapsed -gt 0
353 then
354 next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
355 else
356 next_sample_at=$(($next_sample_at + 1))
358 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
360 printf "\rRewrite $commit ($count/$commits)$progress "
363 git_filter_branch__commit_count=0
365 progress= start_timestamp=
366 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
367 then
368 next_sample_at=0
369 progress="dummy to ensure this is not empty"
370 start_timestamp=$(date '+%s')
373 if test -n "$filter_index" ||
374 test -n "$filter_tree" ||
375 test -n "$filter_subdir"
376 then
377 need_index=t
378 else
379 need_index=
382 eval "$filter_setup" < /dev/null ||
383 die "filter setup failed: $filter_setup"
385 while read commit parents; do
386 git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
388 report_progress
389 test -f "$workdir"/../map/$commit && continue
391 case "$filter_subdir" in
393 if test -n "$need_index"
394 then
395 GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
399 # The commit may not have the subdirectory at all
400 err=$(GIT_ALLOW_NULL_SHA1=1 \
401 git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
402 if ! git rev-parse -q --verify $commit:"$filter_subdir"
403 then
404 rm -f "$GIT_INDEX_FILE"
405 else
406 echo >&2 "$err"
407 false
410 esac || die "Could not initialize the index"
412 GIT_COMMIT=$commit
413 export GIT_COMMIT
414 git cat-file commit "$commit" >../commit ||
415 die "Cannot read commit $commit"
417 eval "$(set_ident <../commit)" ||
418 die "setting author/committer failed for commit $commit"
419 eval "$filter_env" < /dev/null ||
420 die "env filter failed: $filter_env"
422 if [ "$filter_tree" ]; then
423 git checkout-index -f -u -a ||
424 die "Could not checkout the index"
425 # files that $commit removed are now still in the working tree;
426 # remove them, else they would be added again
427 git clean -d -q -f -x
428 eval "$filter_tree" < /dev/null ||
429 die "tree filter failed: $filter_tree"
432 git diff-index -r --name-only --ignore-submodules $commit -- &&
433 git ls-files --others
434 ) > "$tempdir"/tree-state || exit
435 git update-index --add --replace --remove --stdin \
436 < "$tempdir"/tree-state || exit
439 eval "$filter_index" < /dev/null ||
440 die "index filter failed: $filter_index"
442 parentstr=
443 for parent in $parents; do
444 for reparent in $(map "$parent"); do
445 case "$parentstr " in
446 *" -p $reparent "*)
449 parentstr="$parentstr -p $reparent"
451 esac
452 done
453 done
454 if [ "$filter_parent" ]; then
455 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
456 die "parent filter failed: $filter_parent"
460 while IFS='' read -r header_line && test -n "$header_line"
462 # skip header lines...
464 done
465 # and output the actual commit message
467 } <../commit |
468 eval "$filter_msg" > ../message ||
469 die "msg filter failed: $filter_msg"
471 if test -n "$need_index"
472 then
473 tree=$(git write-tree)
474 else
475 tree=$(git rev-parse "$commit^{tree}")
477 workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
478 "$tree" $parentstr < ../message > ../map/$commit ||
479 die "could not write rewritten commit"
480 done <../revs
482 # If we are filtering for paths, as in the case of a subdirectory
483 # filter, it is possible that a specified head is not in the set of
484 # rewritten commits, because it was pruned by the revision walker.
485 # Ancestor remapping fixes this by mapping these heads to the unique
486 # nearest ancestor that survived the pruning.
488 if test "$remap_to_ancestor" = t
489 then
490 while read ref
492 sha1=$(git rev-parse "$ref"^0)
493 test -f "$workdir"/../map/$sha1 && continue
494 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
495 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
496 done < "$tempdir"/heads
499 # Finally update the refs
501 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
502 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
503 echo
504 while read ref
506 # avoid rewriting a ref twice
507 test -f "$orig_namespace$ref" && continue
509 sha1=$(git rev-parse "$ref"^0)
510 rewritten=$(map $sha1)
512 test $sha1 = "$rewritten" &&
513 warn "WARNING: Ref '$ref' is unchanged" &&
514 continue
516 case "$rewritten" in
518 echo "Ref '$ref' was deleted"
519 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
520 die "Could not delete $ref"
522 $_x40)
523 echo "Ref '$ref' was rewritten"
524 if ! git update-ref -m "filter-branch: rewrite" \
525 "$ref" $rewritten $sha1 2>/dev/null; then
526 if test $(git cat-file -t "$ref") = tag; then
527 if test -z "$filter_tag_name"; then
528 warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
529 warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
531 else
532 die "Could not rewrite $ref"
537 # NEEDSWORK: possibly add -Werror, making this an error
538 warn "WARNING: '$ref' was rewritten into multiple commits:"
539 warn "$rewritten"
540 warn "WARNING: Ref '$ref' points to the first one now."
541 rewritten=$(echo "$rewritten" | head -n 1)
542 git update-ref -m "filter-branch: rewrite to first" \
543 "$ref" $rewritten $sha1 ||
544 die "Could not rewrite $ref"
546 esac
547 git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
548 exit
549 done < "$tempdir"/heads
551 # TODO: This should possibly go, with the semantics that all positive given
552 # refs are updated, and their original heads stored in refs/original/
553 # Filter tags
555 if [ "$filter_tag_name" ]; then
556 git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
557 while read sha1 type ref; do
558 ref="${ref#refs/tags/}"
559 # XXX: Rewrite tagged trees as well?
560 if [ "$type" != "commit" -a "$type" != "tag" ]; then
561 continue;
564 if [ "$type" = "tag" ]; then
565 # Dereference to a commit
566 sha1t="$sha1"
567 sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
570 [ -f "../map/$sha1" ] || continue
571 new_sha1="$(cat "../map/$sha1")"
572 GIT_COMMIT="$sha1"
573 export GIT_COMMIT
574 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
575 die "tag name filter failed: $filter_tag_name"
577 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
579 if [ "$type" = "tag" ]; then
580 new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
581 "$new_sha1" "$new_ref"
582 git cat-file tag "$ref" |
583 sed -n \
584 -e '1,/^$/{
585 /^object /d
586 /^type /d
587 /^tag /d
588 }' \
589 -e '/^-----BEGIN PGP SIGNATURE-----/q' \
590 -e 'p' ) |
591 git hash-object -t tag -w --stdin) ||
592 die "Could not create new tag object for $ref"
593 if git cat-file tag "$ref" | \
594 sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
595 then
596 warn "gpg signature stripped from tag object $sha1t"
600 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
601 die "Could not write tag $new_ref"
602 done
605 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
606 unset GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL GIT_AUTHOR_DATE
607 unset GIT_COMMITTER_NAME GIT_COMMITTER_EMAIL GIT_COMMITTER_DATE
608 test -z "$ORIG_GIT_DIR" || {
609 GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
611 test -z "$ORIG_GIT_WORK_TREE" || {
612 GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
613 export GIT_WORK_TREE
615 test -z "$ORIG_GIT_INDEX_FILE" || {
616 GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
617 export GIT_INDEX_FILE
619 test -z "$ORIG_GIT_AUTHOR_NAME" || {
620 GIT_AUTHOR_NAME="$ORIG_GIT_AUTHOR_NAME" &&
621 export GIT_AUTHOR_NAME
623 test -z "$ORIG_GIT_AUTHOR_EMAIL" || {
624 GIT_AUTHOR_EMAIL="$ORIG_GIT_AUTHOR_EMAIL" &&
625 export GIT_AUTHOR_EMAIL
627 test -z "$ORIG_GIT_AUTHOR_DATE" || {
628 GIT_AUTHOR_DATE="$ORIG_GIT_AUTHOR_DATE" &&
629 export GIT_AUTHOR_DATE
631 test -z "$ORIG_GIT_COMMITTER_NAME" || {
632 GIT_COMMITTER_NAME="$ORIG_GIT_COMMITTER_NAME" &&
633 export GIT_COMMITTER_NAME
635 test -z "$ORIG_GIT_COMMITTER_EMAIL" || {
636 GIT_COMMITTER_EMAIL="$ORIG_GIT_COMMITTER_EMAIL" &&
637 export GIT_COMMITTER_EMAIL
639 test -z "$ORIG_GIT_COMMITTER_DATE" || {
640 GIT_COMMITTER_DATE="$ORIG_GIT_COMMITTER_DATE" &&
641 export GIT_COMMITTER_DATE
644 if test -n "$state_branch"
645 then
646 echo "Saving rewrite state to $state_branch" 1>&2
647 state_blob=$(
648 perl -e'opendir D, "../map" or die;
649 open H, "|-", "git hash-object -w --stdin" or die;
650 foreach (sort readdir(D)) {
651 next if m/^\.\.?$/;
652 open F, "<../map/$_" or die;
653 chomp($f = <F>);
654 print H "$_:$f\n" or die;
656 close(H) or die;' || die "Unable to save state")
657 state_tree=$(printf '100644 blob %s\tfilter.map\n' "$state_blob" | git mktree)
658 if test -n "$state_commit"
659 then
660 state_commit=$(echo "Sync" | git commit-tree "$state_tree" -p "$state_commit")
661 else
662 state_commit=$(echo "Sync" | git commit-tree "$state_tree" )
664 git update-ref "$state_branch" "$state_commit"
667 cd "$orig_dir"
668 rm -rf "$tempdir"
670 trap - 0
672 if [ "$(is_bare_repository)" = false ]; then
673 git read-tree -u -m HEAD || exit
676 exit 0