git-filter-branch.sh

   1 #!/bin/sh
   2 #
   3 # Rewrite revision history
   4 # Copyright (c) Petr Baudis, 2006
   5 # Minimal changes to "port" it to core-git (c) Johannes Schindelin, 2007
   6 #
   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.
  10
  11 # The following functions will also be available in the commit filter:
  12
  13 functions=$(cat << \EOF
  14 warn () {
  15         echo "$*" >&2
  16 }
  17
  18 map()
  19 {
  20         # if it was not rewritten, take the original
  21         if test -r "$workdir/../map/$1"
  22         then
  23                 cat "$workdir/../map/$1"
  24         else
  25                 echo "$1"
  26         fi
  27 }
  28
  29 # if you run 'skip_commit "$@"' in a commit filter, it will print
  30 # the (mapped) parents, effectively skipping the commit.
  31
  32 skip_commit()
  33 {
  34         shift;
  35         while [ -n "$1" ];
  36         do
  37                 shift;
  38                 map "$1";
  39                 shift;
  40         done;
  41 }
  42
  43 # if you run 'git_commit_non_empty_tree "$@"' in a commit filter,
  44 # it will skip commits that leave the tree untouched, commit the other.
  45 git_commit_non_empty_tree()
  46 {
  47         if test $# = 3 && test "$1" = $(git rev-parse "$3^{tree}"); then
  48                 map "$3"
  49         elif test $# = 1 && test "$1" = 4b825dc642cb6eb9a060e54bf8d69288fbee4904; then
  50                 :
  51         else
  52                 git commit-tree "$@"
  53         fi
  54 }
  55 # override die(): this version puts in an extra line break, so that
  56 # the progress is still visible
  57
  58 die()
  59 {
  60         echo >&2
  61         echo "$*" >&2
  62         exit 1
  63 }
  64 EOF
  65 )
  66
  67 eval "$functions"
  68
  69 finish_ident() {
  70         # Ensure non-empty id name.
  71         echo "case \"\$GIT_$1_NAME\" in \"\") GIT_$1_NAME=\"\${GIT_$1_EMAIL%%@*}\" && export GIT_$1_NAME;; esac"
  72         # And make sure everything is exported.
  73         echo "export GIT_$1_NAME"
  74         echo "export GIT_$1_EMAIL"
  75         echo "export GIT_$1_DATE"
  76 }
  77
  78 set_ident () {
  79         parse_ident_from_commit author AUTHOR committer COMMITTER
  80         finish_ident AUTHOR
  81         finish_ident COMMITTER
  82 }
  83
  84 USAGE="[--setup <command>] [--subdirectory-filter <directory>] [--env-filter <command>]
  85         [--tree-filter <command>] [--index-filter <command>]
  86         [--parent-filter <command>] [--msg-filter <command>]
  87         [--commit-filter <command>] [--tag-name-filter <command>]
  88         [--original <namespace>]
  89         [-d <directory>] [-f | --force] [--state-branch <branch>]
  90         [--] [<rev-list options>...]"
  91
  92 OPTIONS_SPEC=
  93 . git-sh-setup
  94
  95 if [ "$(is_bare_repository)" = false ]; then
  96         require_clean_work_tree 'rewrite branches'
  97 fi
  98
  99 tempdir=.git-rewrite
 100 filter_setup=
 101 filter_env=
 102 filter_tree=
 103 filter_index=
 104 filter_parent=
 105 filter_msg=cat
 106 filter_commit=
 107 filter_tag_name=
 108 filter_subdir=
 109 state_branch=
 110 orig_namespace=refs/original/
 111 force=
 112 prune_empty=
 113 remap_to_ancestor=
 114 while :
 115 do
 116         case "$1" in
 117         --)
 118                 shift
 119                 break
 120                 ;;
 121         --force|-f)
 122                 shift
 123                 force=t
 124                 continue
 125                 ;;
 126         --remap-to-ancestor)
 127                 # deprecated ($remap_to_ancestor is set now automatically)
 128                 shift
 129                 remap_to_ancestor=t
 130                 continue
 131                 ;;
 132         --prune-empty)
 133                 shift
 134                 prune_empty=t
 135                 continue
 136                 ;;
 137         -*)
 138                 ;;
 139         *)
 140                 break;
 141         esac
 142
 143         # all switches take one argument
 144         ARG="$1"
 145         case "$#" in 1) usage ;; esac
 146         shift
 147         OPTARG="$1"
 148         shift
 149
 150         case "$ARG" in
 151         -d)
 152                 tempdir="$OPTARG"
 153                 ;;
 154         --setup)
 155                 filter_setup="$OPTARG"
 156                 ;;
 157         --subdirectory-filter)
 158                 filter_subdir="$OPTARG"
 159                 remap_to_ancestor=t
 160                 ;;
 161         --env-filter)
 162                 filter_env="$OPTARG"
 163                 ;;
 164         --tree-filter)
 165                 filter_tree="$OPTARG"
 166                 ;;
 167         --index-filter)
 168                 filter_index="$OPTARG"
 169                 ;;
 170         --parent-filter)
 171                 filter_parent="$OPTARG"
 172                 ;;
 173         --msg-filter)
 174                 filter_msg="$OPTARG"
 175                 ;;
 176         --commit-filter)
 177                 filter_commit="$functions; $OPTARG"
 178                 ;;
 179         --tag-name-filter)
 180                 filter_tag_name="$OPTARG"
 181                 ;;
 182         --original)
 183                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 184                 ;;
 185         --state-branch)
 186                 state_branch="$OPTARG"
 187                 ;;
 188         *)
 189                 usage
 190                 ;;
 191         esac
 192 done
 193
 194 case "$prune_empty,$filter_commit" in
 195 ,)
 196         filter_commit='git commit-tree "$@"';;
 197 t,)
 198         filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
 199 ,*)
 200         ;;
 201 *)
 202         die "Cannot set --prune-empty and --commit-filter at the same time"
 203 esac
 204
 205 case "$force" in
 206 t)
 207         rm -rf "$tempdir"
 208 ;;
 209 '')
 210         test -d "$tempdir" &&
 211                 die "$tempdir already exists, please remove it"
 212 esac
 213 orig_dir=$(pwd)
 214 mkdir -p "$tempdir/t" &&
 215 tempdir="$(cd "$tempdir"; pwd)" &&
 216 cd "$tempdir/t" &&
 217 workdir="$(pwd)" ||
 218 die ""
 219
 220 # Remove tempdir on exit
 221 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
 222
 223 ORIG_GIT_DIR="$GIT_DIR"
 224 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 225 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 226 ORIG_GIT_AUTHOR_NAME="$GIT_AUTHOR_NAME"
 227 ORIG_GIT_AUTHOR_EMAIL="$GIT_AUTHOR_EMAIL"
 228 ORIG_GIT_AUTHOR_DATE="$GIT_AUTHOR_DATE"
 229 ORIG_GIT_COMMITTER_NAME="$GIT_COMMITTER_NAME"
 230 ORIG_GIT_COMMITTER_EMAIL="$GIT_COMMITTER_EMAIL"
 231 ORIG_GIT_COMMITTER_DATE="$GIT_COMMITTER_DATE"
 232
 233 GIT_WORK_TREE=.
 234 export GIT_DIR GIT_WORK_TREE
 235
 236 # Make sure refs/original is empty
 237 git for-each-ref > "$tempdir"/backup-refs || exit
 238 while read sha1 type name
 239 do
 240         case "$force,$name" in
 241         ,$orig_namespace*)
 242                 die "Cannot create a new backup.
 243 A previous backup already exists in $orig_namespace
 244 Force overwriting the backup with -f"
 245         ;;
 246         t,$orig_namespace*)
 247                 git update-ref -d "$name" $sha1
 248         ;;
 249         esac
 250 done < "$tempdir"/backup-refs
 251
 252 # The refs should be updated if their heads were rewritten
 253 git rev-parse --no-flags --revs-only --symbolic-full-name \
 254         --default HEAD "$@" > "$tempdir"/raw-refs || exit
 255 while read ref
 256 do
 257         case "$ref" in ^?*) continue ;; esac
 258
 259         if git rev-parse --verify "$ref"^0 >/dev/null 2>&1
 260         then
 261                 echo "$ref"
 262         else
 263                 warn "WARNING: not rewriting '$ref' (not a committish)"
 264         fi
 265 done >"$tempdir"/heads <"$tempdir"/raw-refs
 266
 267 test -s "$tempdir"/heads ||
 268         die "You must specify a ref to rewrite."
 269
 270 GIT_INDEX_FILE="$(pwd)/../index"
 271 export GIT_INDEX_FILE
 272
 273 # map old->new commit ids for rewriting parents
 274 mkdir ../map || die "Could not create map/ directory"
 275
 276 if test -n "$state_branch"
 277 then
 278         state_commit=$(git rev-parse --no-flags --revs-only "$state_branch")
 279         if test -n "$state_commit"
 280         then
 281                 echo "Populating map from $state_branch ($state_commit)" 1>&2
 282                 perl -e'open(MAP, "-|", "git show $ARGV[0]:filter.map") or die;
 283                         while (<MAP>) {
 284                                 m/(.*):(.*)/ or die;
 285                                 open F, ">../map/$1" or die;
 286                                 print F "$2" or die;
 287                                 close(F) or die;
 288                         }
 289                         close(MAP) or die;' "$state_commit" \
 290                                 || die "Unable to load state from $state_branch:filter.map"
 291         else
 292                 echo "Branch $state_branch does not exist. Will create" 1>&2
 293         fi
 294 fi
 295
 296 # we need "--" only if there are no path arguments in $@
 297 nonrevs=$(git rev-parse --no-revs "$@") || exit
 298 if test -z "$nonrevs"
 299 then
 300         dashdash=--
 301 else
 302         dashdash=
 303         remap_to_ancestor=t
 304 fi
 305
 306 git rev-parse --revs-only "$@" >../parse
 307
 308 case "$filter_subdir" in
 309 "")
 310         eval set -- "$(git rev-parse --sq --no-revs "$@")"
 311         ;;
 312 *)
 313         eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
 314                 "$filter_subdir")"
 315         ;;
 316 esac
 317
 318 git rev-list --reverse --topo-order --default HEAD \
 319         --parents --simplify-merges --stdin "$@" <../parse >../revs ||
 320         die "Could not get the commits"
 321 commits=$(wc -l <../revs | tr -d " ")
 322
 323 test $commits -eq 0 && die_with_status 2 "Found nothing to rewrite"
 324
 325 # Rewrite the commits
 326 report_progress ()
 327 {
 328         if test -n "$progress" &&
 329                 test $git_filter_branch__commit_count -gt $next_sample_at
 330         then
 331                 count=$git_filter_branch__commit_count
 332
 333                 now=$(date +%s)
 334                 elapsed=$(($now - $start_timestamp))
 335                 remaining=$(( ($commits - $count) * $elapsed / $count ))
 336                 if test $elapsed -gt 0
 337                 then
 338                         next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
 339                 else
 340                         next_sample_at=$(($next_sample_at + 1))
 341                 fi
 342                 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
 343         fi
 344         printf "\rRewrite $commit ($count/$commits)$progress    "
 345 }
 346
 347 git_filter_branch__commit_count=0
 348
 349 progress= start_timestamp=
 350 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
 351 then
 352         next_sample_at=0
 353         progress="dummy to ensure this is not empty"
 354         start_timestamp=$(date '+%s')
 355 fi
 356
 357 if test -n "$filter_index" ||
 358    test -n "$filter_tree" ||
 359    test -n "$filter_subdir"
 360 then
 361         need_index=t
 362 else
 363         need_index=
 364 fi
 365
 366 eval "$filter_setup" < /dev/null ||
 367         die "filter setup failed: $filter_setup"
 368
 369 while read commit parents; do
 370         git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
 371
 372         report_progress
 373
 374         case "$filter_subdir" in
 375         "")
 376                 if test -n "$need_index"
 377                 then
 378                         GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
 379                 fi
 380                 ;;
 381         *)
 382                 # The commit may not have the subdirectory at all
 383                 err=$(GIT_ALLOW_NULL_SHA1=1 \
 384                       git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 385                         if ! git rev-parse -q --verify $commit:"$filter_subdir"
 386                         then
 387                                 rm -f "$GIT_INDEX_FILE"
 388                         else
 389                                 echo >&2 "$err"
 390                                 false
 391                         fi
 392                 }
 393         esac || die "Could not initialize the index"
 394
 395         GIT_COMMIT=$commit
 396         export GIT_COMMIT
 397         git cat-file commit "$commit" >../commit ||
 398                 die "Cannot read commit $commit"
 399
 400         eval "$(set_ident <../commit)" ||
 401                 die "setting author/committer failed for commit $commit"
 402         eval "$filter_env" < /dev/null ||
 403                 die "env filter failed: $filter_env"
 404
 405         if [ "$filter_tree" ]; then
 406                 git checkout-index -f -u -a ||
 407                         die "Could not checkout the index"
 408                 # files that $commit removed are now still in the working tree;
 409                 # remove them, else they would be added again
 410                 git clean -d -q -f -x
 411                 eval "$filter_tree" < /dev/null ||
 412                         die "tree filter failed: $filter_tree"
 413
 414                 (
 415                         git diff-index -r --name-only --ignore-submodules $commit -- &&
 416                         git ls-files --others
 417                 ) > "$tempdir"/tree-state || exit
 418                 git update-index --add --replace --remove --stdin \
 419                         < "$tempdir"/tree-state || exit
 420         fi
 421
 422         eval "$filter_index" < /dev/null ||
 423                 die "index filter failed: $filter_index"
 424
 425         parentstr=
 426         for parent in $parents; do
 427                 for reparent in $(map "$parent"); do
 428                         case "$parentstr " in
 429                         *" -p $reparent "*)
 430                                 ;;
 431                         *)
 432                                 parentstr="$parentstr -p $reparent"
 433                                 ;;
 434                         esac
 435                 done
 436         done
 437         if [ "$filter_parent" ]; then
 438                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 439                                 die "parent filter failed: $filter_parent"
 440         fi
 441
 442         {
 443                 while IFS='' read -r header_line && test -n "$header_line"
 444                 do
 445                         # skip header lines...
 446                         :;
 447                 done
 448                 # and output the actual commit message
 449                 cat
 450         } <../commit |
 451                 eval "$filter_msg" > ../message ||
 452                         die "msg filter failed: $filter_msg"
 453
 454         if test -n "$need_index"
 455         then
 456                 tree=$(git write-tree)
 457         else
 458                 tree=$(git rev-parse "$commit^{tree}")
 459         fi
 460         workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
 461                 "$tree" $parentstr < ../message > ../map/$commit ||
 462                         die "could not write rewritten commit"
 463 done <../revs
 464
 465 # If we are filtering for paths, as in the case of a subdirectory
 466 # filter, it is possible that a specified head is not in the set of
 467 # rewritten commits, because it was pruned by the revision walker.
 468 # Ancestor remapping fixes this by mapping these heads to the unique
 469 # nearest ancestor that survived the pruning.
 470
 471 if test "$remap_to_ancestor" = t
 472 then
 473         while read ref
 474         do
 475                 sha1=$(git rev-parse "$ref"^0)
 476                 test -f "$workdir"/../map/$sha1 && continue
 477                 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
 478                 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
 479         done < "$tempdir"/heads
 480 fi
 481
 482 # Finally update the refs
 483
 484 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 485 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 486 echo
 487 while read ref
 488 do
 489         # avoid rewriting a ref twice
 490         test -f "$orig_namespace$ref" && continue
 491
 492         sha1=$(git rev-parse "$ref"^0)
 493         rewritten=$(map $sha1)
 494
 495         test $sha1 = "$rewritten" &&
 496                 warn "WARNING: Ref '$ref' is unchanged" &&
 497                 continue
 498
 499         case "$rewritten" in
 500         '')
 501                 echo "Ref '$ref' was deleted"
 502                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 503                         die "Could not delete $ref"
 504         ;;
 505         $_x40)
 506                 echo "Ref '$ref' was rewritten"
 507                 if ! git update-ref -m "filter-branch: rewrite" \
 508                                         "$ref" $rewritten $sha1 2>/dev/null; then
 509                         if test $(git cat-file -t "$ref") = tag; then
 510                                 if test -z "$filter_tag_name"; then
 511                                         warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
 512                                         warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
 513                                 fi
 514                         else
 515                                 die "Could not rewrite $ref"
 516                         fi
 517                 fi
 518         ;;
 519         *)
 520                 # NEEDSWORK: possibly add -Werror, making this an error
 521                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 522                 warn "$rewritten"
 523                 warn "WARNING: Ref '$ref' points to the first one now."
 524                 rewritten=$(echo "$rewritten" | head -n 1)
 525                 git update-ref -m "filter-branch: rewrite to first" \
 526                                 "$ref" $rewritten $sha1 ||
 527                         die "Could not rewrite $ref"
 528         ;;
 529         esac
 530         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
 531                  exit
 532 done < "$tempdir"/heads
 533
 534 # TODO: This should possibly go, with the semantics that all positive given
 535 #       refs are updated, and their original heads stored in refs/original/
 536 # Filter tags
 537
 538 if [ "$filter_tag_name" ]; then
 539         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 540         while read sha1 type ref; do
 541                 ref="${ref#refs/tags/}"
 542                 # XXX: Rewrite tagged trees as well?
 543                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 544                         continue;
 545                 fi
 546
 547                 if [ "$type" = "tag" ]; then
 548                         # Dereference to a commit
 549                         sha1t="$sha1"
 550                         sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
 551                 fi
 552
 553                 [ -f "../map/$sha1" ] || continue
 554                 new_sha1="$(cat "../map/$sha1")"
 555                 GIT_COMMIT="$sha1"
 556                 export GIT_COMMIT
 557                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 558                         die "tag name filter failed: $filter_tag_name"
 559
 560                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 561
 562                 if [ "$type" = "tag" ]; then
 563                         new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
 564                                                 "$new_sha1" "$new_ref"
 565                                 git cat-file tag "$ref" |
 566                                 sed -n \
 567                                     -e '1,/^$/{
 568                                           /^object /d
 569                                           /^type /d
 570                                           /^tag /d
 571                                         }' \
 572                                     -e '/^-----BEGIN PGP SIGNATURE-----/q' \
 573                                     -e 'p' ) |
 574                                 git hash-object -t tag -w --stdin) ||
 575                                 die "Could not create new tag object for $ref"
 576                         if git cat-file tag "$ref" | \
 577                            sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
 578                         then
 579                                 warn "gpg signature stripped from tag object $sha1t"
 580                         fi
 581                 fi
 582
 583                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 584                         die "Could not write tag $new_ref"
 585         done
 586 fi
 587
 588 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 589 unset GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL GIT_AUTHOR_DATE
 590 unset GIT_COMMITTER_NAME GIT_COMMITTER_EMAIL GIT_COMMITTER_DATE
 591 test -z "$ORIG_GIT_DIR" || {
 592         GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 593 }
 594 test -z "$ORIG_GIT_WORK_TREE" || {
 595         GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 596         export GIT_WORK_TREE
 597 }
 598 test -z "$ORIG_GIT_INDEX_FILE" || {
 599         GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 600         export GIT_INDEX_FILE
 601 }
 602 test -z "$ORIG_GIT_AUTHOR_NAME" || {
 603         GIT_AUTHOR_NAME="$ORIG_GIT_AUTHOR_NAME" &&
 604         export GIT_AUTHOR_NAME
 605 }
 606 test -z "$ORIG_GIT_AUTHOR_EMAIL" || {
 607         GIT_AUTHOR_EMAIL="$ORIG_GIT_AUTHOR_EMAIL" &&
 608         export GIT_AUTHOR_EMAIL
 609 }
 610 test -z "$ORIG_GIT_AUTHOR_DATE" || {
 611         GIT_AUTHOR_DATE="$ORIG_GIT_AUTHOR_DATE" &&
 612         export GIT_AUTHOR_DATE
 613 }
 614 test -z "$ORIG_GIT_COMMITTER_NAME" || {
 615         GIT_COMMITTER_NAME="$ORIG_GIT_COMMITTER_NAME" &&
 616         export GIT_COMMITTER_NAME
 617 }
 618 test -z "$ORIG_GIT_COMMITTER_EMAIL" || {
 619         GIT_COMMITTER_EMAIL="$ORIG_GIT_COMMITTER_EMAIL" &&
 620         export GIT_COMMITTER_EMAIL
 621 }
 622 test -z "$ORIG_GIT_COMMITTER_DATE" || {
 623         GIT_COMMITTER_DATE="$ORIG_GIT_COMMITTER_DATE" &&
 624         export GIT_COMMITTER_DATE
 625 }
 626
 627 if test -n "$state_branch"
 628 then
 629         echo "Saving rewrite state to $state_branch" 1>&2
 630         state_blob=$(
 631                 perl -e'opendir D, "../map" or die;
 632                         open H, "|-", "git hash-object -w --stdin" or die;
 633                         foreach (sort readdir(D)) {
 634                                 next if m/^\.\.?$/;
 635                                 open F, "<../map/$_" or die;
 636                                 chomp($f = <F>);
 637                                 print H "$_:$f\n" or die;
 638                         }
 639                         close(H) or die;' || die "Unable to save state")
 640         state_tree=$(printf '100644 blob %s\tfilter.map\n' "$state_blob" | git mktree)
 641         if test -n "$state_commit"
 642         then
 643                 state_commit=$(echo "Sync" | git commit-tree "$state_tree" -p "$state_commit")
 644         else
 645                 state_commit=$(echo "Sync" | git commit-tree "$state_tree" )
 646         fi
 647         git update-ref "$state_branch" "$state_commit"
 648 fi
 649
 650 cd "$orig_dir"
 651 rm -rf "$tempdir"
 652
 653 trap - 0
 654
 655 if [ "$(is_bare_repository)" = false ]; then
 656         git read-tree -u -m HEAD || exit
 657 fi
 658
 659 exit 0