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_$1_DATE"
  76 }
  77
  78 set_ident () {
  79         parse_ident_from_commit author AUTHOR committer COMMITTER
  80         finish_ident AUTHOR
  81         finish_ident COMMITTER
  82 }
  83
  84 USAGE="[--setup <command>] [--env-filter <command>]
  85         [--tree-filter <command>] [--index-filter <command>]
  86         [--parent-filter <command>] [--msg-filter <command>]
  87         [--commit-filter <command>] [--tag-name-filter <command>]
  88         [--subdirectory-filter <directory>] [--original <namespace>]
  89         [-d <directory>] [-f | --force] [--state-branch <branch>]
  90         [--] [<rev-list options>...]"
  91
  92 OPTIONS_SPEC=
  93 . git-sh-setup
  94
  95 if [ "$(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         --env-filter)
 158                 filter_env="$OPTARG"
 159                 ;;
 160         --tree-filter)
 161                 filter_tree="$OPTARG"
 162                 ;;
 163         --index-filter)
 164                 filter_index="$OPTARG"
 165                 ;;
 166         --parent-filter)
 167                 filter_parent="$OPTARG"
 168                 ;;
 169         --msg-filter)
 170                 filter_msg="$OPTARG"
 171                 ;;
 172         --commit-filter)
 173                 filter_commit="$functions; $OPTARG"
 174                 ;;
 175         --tag-name-filter)
 176                 filter_tag_name="$OPTARG"
 177                 ;;
 178         --subdirectory-filter)
 179                 filter_subdir="$OPTARG"
 180                 remap_to_ancestor=t
 181                 ;;
 182         --original)
 183                 orig_namespace=$(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-heads || exit
 255 sed -e '/^^/d' "$tempdir"/raw-heads >"$tempdir"/heads
 256
 257 test -s "$tempdir"/heads ||
 258         die "You must specify a ref to rewrite."
 259
 260 GIT_INDEX_FILE="$(pwd)/../index"
 261 export GIT_INDEX_FILE
 262
 263 # map old->new commit ids for rewriting parents
 264 mkdir ../map || die "Could not create map/ directory"
 265
 266 if test -n "$state_branch"
 267 then
 268         state_commit=$(git rev-parse --no-flags --revs-only "$state_branch")
 269         if test -n "$state_commit"
 270         then
 271                 echo "Populating map from $state_branch ($state_commit)" 1>&2
 272                 perl -e'open(MAP, "-|", "git show $ARGV[0]:filter.map") or die;
 273                         while (<MAP>) {
 274                                 m/(.*):(.*)/ or die;
 275                                 open F, ">../map/$1" or die;
 276                                 print F "$2" or die;
 277                                 close(F) or die;
 278                         }
 279                         close(MAP) or die;' "$state_commit" \
 280                                 || die "Unable to load state from $state_branch:filter.map"
 281         else
 282                 echo "Branch $state_branch does not exist. Will create" 1>&2
 283         fi
 284 fi
 285
 286 # we need "--" only if there are no path arguments in $@
 287 nonrevs=$(git rev-parse --no-revs "$@") || exit
 288 if test -z "$nonrevs"
 289 then
 290         dashdash=--
 291 else
 292         dashdash=
 293         remap_to_ancestor=t
 294 fi
 295
 296 git rev-parse --revs-only "$@" >../parse
 297
 298 case "$filter_subdir" in
 299 "")
 300         eval set -- "$(git rev-parse --sq --no-revs "$@")"
 301         ;;
 302 *)
 303         eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
 304                 "$filter_subdir")"
 305         ;;
 306 esac
 307
 308 git rev-list --reverse --topo-order --default HEAD \
 309         --parents --simplify-merges --stdin "$@" <../parse >../revs ||
 310         die "Could not get the commits"
 311 commits=$(wc -l <../revs | tr -d " ")
 312
 313 test $commits -eq 0 && die "Found nothing to rewrite"
 314
 315 # Rewrite the commits
 316 report_progress ()
 317 {
 318         if test -n "$progress" &&
 319                 test $git_filter_branch__commit_count -gt $next_sample_at
 320         then
 321                 count=$git_filter_branch__commit_count
 322
 323                 now=$(date +%s)
 324                 elapsed=$(($now - $start_timestamp))
 325                 remaining=$(( ($commits - $count) * $elapsed / $count ))
 326                 if test $elapsed -gt 0
 327                 then
 328                         next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
 329                 else
 330                         next_sample_at=$(($next_sample_at + 1))
 331                 fi
 332                 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
 333         fi
 334         printf "\rRewrite $commit ($count/$commits)$progress    "
 335 }
 336
 337 git_filter_branch__commit_count=0
 338
 339 progress= start_timestamp=
 340 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
 341 then
 342         next_sample_at=0
 343         progress="dummy to ensure this is not empty"
 344         start_timestamp=$(date '+%s')
 345 fi
 346
 347 if test -n "$filter_index" ||
 348    test -n "$filter_tree" ||
 349    test -n "$filter_subdir"
 350 then
 351         need_index=t
 352 else
 353         need_index=
 354 fi
 355
 356 eval "$filter_setup" < /dev/null ||
 357         die "filter setup failed: $filter_setup"
 358
 359 while read commit parents; do
 360         git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
 361
 362         report_progress
 363
 364         case "$filter_subdir" in
 365         "")
 366                 if test -n "$need_index"
 367                 then
 368                         GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
 369                 fi
 370                 ;;
 371         *)
 372                 # The commit may not have the subdirectory at all
 373                 err=$(GIT_ALLOW_NULL_SHA1=1 \
 374                       git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 375                         if ! git rev-parse -q --verify $commit:"$filter_subdir"
 376                         then
 377                                 rm -f "$GIT_INDEX_FILE"
 378                         else
 379                                 echo >&2 "$err"
 380                                 false
 381                         fi
 382                 }
 383         esac || die "Could not initialize the index"
 384
 385         GIT_COMMIT=$commit
 386         export GIT_COMMIT
 387         git cat-file commit "$commit" >../commit ||
 388                 die "Cannot read commit $commit"
 389
 390         eval "$(set_ident <../commit)" ||
 391                 die "setting author/committer failed for commit $commit"
 392         eval "$filter_env" < /dev/null ||
 393                 die "env filter failed: $filter_env"
 394
 395         if [ "$filter_tree" ]; then
 396                 git checkout-index -f -u -a ||
 397                         die "Could not checkout the index"
 398                 # files that $commit removed are now still in the working tree;
 399                 # remove them, else they would be added again
 400                 git clean -d -q -f -x
 401                 eval "$filter_tree" < /dev/null ||
 402                         die "tree filter failed: $filter_tree"
 403
 404                 (
 405                         git diff-index -r --name-only --ignore-submodules $commit -- &&
 406                         git ls-files --others
 407                 ) > "$tempdir"/tree-state || exit
 408                 git update-index --add --replace --remove --stdin \
 409                         < "$tempdir"/tree-state || exit
 410         fi
 411
 412         eval "$filter_index" < /dev/null ||
 413                 die "index filter failed: $filter_index"
 414
 415         parentstr=
 416         for parent in $parents; do
 417                 for reparent in $(map "$parent"); do
 418                         case "$parentstr " in
 419                         *" -p $reparent "*)
 420                                 ;;
 421                         *)
 422                                 parentstr="$parentstr -p $reparent"
 423                                 ;;
 424                         esac
 425                 done
 426         done
 427         if [ "$filter_parent" ]; then
 428                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 429                                 die "parent filter failed: $filter_parent"
 430         fi
 431
 432         {
 433                 while IFS='' read -r header_line && test -n "$header_line"
 434                 do
 435                         # skip header lines...
 436                         :;
 437                 done
 438                 # and output the actual commit message
 439                 cat
 440         } <../commit |
 441                 eval "$filter_msg" > ../message ||
 442                         die "msg filter failed: $filter_msg"
 443
 444         if test -n "$need_index"
 445         then
 446                 tree=$(git write-tree)
 447         else
 448                 tree=$(git rev-parse "$commit^{tree}")
 449         fi
 450         workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
 451                 "$tree" $parentstr < ../message > ../map/$commit ||
 452                         die "could not write rewritten commit"
 453 done <../revs
 454
 455 # If we are filtering for paths, as in the case of a subdirectory
 456 # filter, it is possible that a specified head is not in the set of
 457 # rewritten commits, because it was pruned by the revision walker.
 458 # Ancestor remapping fixes this by mapping these heads to the unique
 459 # nearest ancestor that survived the pruning.
 460
 461 if test "$remap_to_ancestor" = t
 462 then
 463         while read ref
 464         do
 465                 sha1=$(git rev-parse "$ref"^0)
 466                 test -f "$workdir"/../map/$sha1 && continue
 467                 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
 468                 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
 469         done < "$tempdir"/heads
 470 fi
 471
 472 # Finally update the refs
 473
 474 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 475 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 476 echo
 477 while read ref
 478 do
 479         # avoid rewriting a ref twice
 480         test -f "$orig_namespace$ref" && continue
 481
 482         sha1=$(git rev-parse "$ref"^0)
 483         rewritten=$(map $sha1)
 484
 485         test $sha1 = "$rewritten" &&
 486                 warn "WARNING: Ref '$ref' is unchanged" &&
 487                 continue
 488
 489         case "$rewritten" in
 490         '')
 491                 echo "Ref '$ref' was deleted"
 492                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 493                         die "Could not delete $ref"
 494         ;;
 495         $_x40)
 496                 echo "Ref '$ref' was rewritten"
 497                 if ! git update-ref -m "filter-branch: rewrite" \
 498                                         "$ref" $rewritten $sha1 2>/dev/null; then
 499                         if test $(git cat-file -t "$ref") = tag; then
 500                                 if test -z "$filter_tag_name"; then
 501                                         warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
 502                                         warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
 503                                 fi
 504                         else
 505                                 die "Could not rewrite $ref"
 506                         fi
 507                 fi
 508         ;;
 509         *)
 510                 # NEEDSWORK: possibly add -Werror, making this an error
 511                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 512                 warn "$rewritten"
 513                 warn "WARNING: Ref '$ref' points to the first one now."
 514                 rewritten=$(echo "$rewritten" | head -n 1)
 515                 git update-ref -m "filter-branch: rewrite to first" \
 516                                 "$ref" $rewritten $sha1 ||
 517                         die "Could not rewrite $ref"
 518         ;;
 519         esac
 520         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
 521                  exit
 522 done < "$tempdir"/heads
 523
 524 # TODO: This should possibly go, with the semantics that all positive given
 525 #       refs are updated, and their original heads stored in refs/original/
 526 # Filter tags
 527
 528 if [ "$filter_tag_name" ]; then
 529         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 530         while read sha1 type ref; do
 531                 ref="${ref#refs/tags/}"
 532                 # XXX: Rewrite tagged trees as well?
 533                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 534                         continue;
 535                 fi
 536
 537                 if [ "$type" = "tag" ]; then
 538                         # Dereference to a commit
 539                         sha1t="$sha1"
 540                         sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
 541                 fi
 542
 543                 [ -f "../map/$sha1" ] || continue
 544                 new_sha1="$(cat "../map/$sha1")"
 545                 GIT_COMMIT="$sha1"
 546                 export GIT_COMMIT
 547                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 548                         die "tag name filter failed: $filter_tag_name"
 549
 550                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 551
 552                 if [ "$type" = "tag" ]; then
 553                         new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
 554                                                 "$new_sha1" "$new_ref"
 555                                 git cat-file tag "$ref" |
 556                                 sed -n \
 557                                     -e '1,/^$/{
 558                                           /^object /d
 559                                           /^type /d
 560                                           /^tag /d
 561                                         }' \
 562                                     -e '/^-----BEGIN PGP SIGNATURE-----/q' \
 563                                     -e 'p' ) |
 564                                 git hash-object -t tag -w --stdin) ||
 565                                 die "Could not create new tag object for $ref"
 566                         if git cat-file tag "$ref" | \
 567                            sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
 568                         then
 569                                 warn "gpg signature stripped from tag object $sha1t"
 570                         fi
 571                 fi
 572
 573                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 574                         die "Could not write tag $new_ref"
 575         done
 576 fi
 577
 578 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 579 unset GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL GIT_AUTHOR_DATE
 580 unset GIT_COMMITTER_NAME GIT_COMMITTER_EMAIL GIT_COMMITTER_DATE
 581 test -z "$ORIG_GIT_DIR" || {
 582         GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 583 }
 584 test -z "$ORIG_GIT_WORK_TREE" || {
 585         GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 586         export GIT_WORK_TREE
 587 }
 588 test -z "$ORIG_GIT_INDEX_FILE" || {
 589         GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 590         export GIT_INDEX_FILE
 591 }
 592 test -z "$ORIG_GIT_AUTHOR_NAME" || {
 593         GIT_AUTHOR_NAME="$ORIG_GIT_AUTHOR_NAME" &&
 594         export GIT_AUTHOR_NAME
 595 }
 596 test -z "$ORIG_GIT_AUTHOR_EMAIL" || {
 597         GIT_AUTHOR_EMAIL="$ORIG_GIT_AUTHOR_EMAIL" &&
 598         export GIT_AUTHOR_EMAIL
 599 }
 600 test -z "$ORIG_GIT_AUTHOR_DATE" || {
 601         GIT_AUTHOR_DATE="$ORIG_GIT_AUTHOR_DATE" &&
 602         export GIT_AUTHOR_DATE
 603 }
 604 test -z "$ORIG_GIT_COMMITTER_NAME" || {
 605         GIT_COMMITTER_NAME="$ORIG_GIT_COMMITTER_NAME" &&
 606         export GIT_COMMITTER_NAME
 607 }
 608 test -z "$ORIG_GIT_COMMITTER_EMAIL" || {
 609         GIT_COMMITTER_EMAIL="$ORIG_GIT_COMMITTER_EMAIL" &&
 610         export GIT_COMMITTER_EMAIL
 611 }
 612 test -z "$ORIG_GIT_COMMITTER_DATE" || {
 613         GIT_COMMITTER_DATE="$ORIG_GIT_COMMITTER_DATE" &&
 614         export GIT_COMMITTER_DATE
 615 }
 616
 617 if test -n "$state_branch"
 618 then
 619         echo "Saving rewrite state to $state_branch" 1>&2
 620         state_blob=$(
 621                 perl -e'opendir D, "../map" or die;
 622                         open H, "|-", "git hash-object -w --stdin" or die;
 623                         foreach (sort readdir(D)) {
 624                                 next if m/^\.\.?$/;
 625                                 open F, "<../map/$_" or die;
 626                                 chomp($f = <F>);
 627                                 print H "$_:$f\n" or die;
 628                         }
 629                         close(H) or die;' || die "Unable to save state")
 630         state_tree=$(/bin/echo -e "100644 blob $state_blob\tfilter.map" | git mktree)
 631         if test -n "$state_commit"
 632         then
 633                 state_commit=$(/bin/echo "Sync" | git commit-tree "$state_tree" -p "$state_commit")
 634         else
 635                 state_commit=$(/bin/echo "Sync" | git commit-tree "$state_tree" )
 636         fi
 637         git update-ref "$state_branch" "$state_commit"
 638 fi
 639
 640 cd "$orig_dir"
 641 rm -rf "$tempdir"
 642
 643 trap - 0
 644
 645 if [ "$(is_bare_repository)" = false ]; then
 646         git read-tree -u -m HEAD || exit
 647 fi
 648
 649 exit 0