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=$(cat << \EOF
  14 EMPTY_TREE=$(git hash-object -t tree /dev/null)
  15
  16 warn () {
  17         echo "$*" >&2
  18 }
  19
  20 map()
  21 {
  22         # if it was not rewritten, take the original
  23         if test -r "$workdir/../map/$1"
  24         then
  25                 cat "$workdir/../map/$1"
  26         else
  27                 echo "$1"
  28         fi
  29 }
  30
  31 # if you run 'skip_commit "$@"' in a commit filter, it will print
  32 # the (mapped) parents, effectively skipping the commit.
  33
  34 skip_commit()
  35 {
  36         shift;
  37         while [ -n "$1" ];
  38         do
  39                 shift;
  40                 map "$1";
  41                 shift;
  42         done;
  43 }
  44
  45 # if you run 'git_commit_non_empty_tree "$@"' in a commit filter,
  46 # it will skip commits that leave the tree untouched, commit the other.
  47 git_commit_non_empty_tree()
  48 {
  49         if test $# = 3 && test "$1" = $(git rev-parse "$3^{tree}"); then
  50                 map "$3"
  51         elif test $# = 1 && test "$1" = $EMPTY_TREE; then
  52                 :
  53         else
  54                 git commit-tree "$@"
  55         fi
  56 }
  57 # override die(): this version puts in an extra line break, so that
  58 # the progress is still visible
  59
  60 die()
  61 {
  62         echo >&2
  63         echo "$*" >&2
  64         exit 1
  65 }
  66 EOF
  67 )
  68
  69 eval "$functions"
  70
  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_$1_DATE"
  78 }
  79
  80 set_ident () {
  81         parse_ident_from_commit author AUTHOR committer COMMITTER
  82         finish_ident AUTHOR
  83         finish_ident COMMITTER
  84 }
  85
  86 if test -z "$FILTER_BRANCH_SQUELCH_WARNING$GIT_TEST_DISALLOW_ABBREVIATED_OPTIONS"
  87 then
  88         cat <<EOF
  89 WARNING: git-filter-branch has a glut of gotchas generating mangled history
  90          rewrites.  Hit Ctrl-C before proceeding to abort, then use an
  91          alternative filtering tool such as 'git filter-repo'
  92          (https://github.com/newren/git-filter-repo/) instead.  See the
  93          filter-branch manual page for more details; to squelch this warning,
  94          set FILTER_BRANCH_SQUELCH_WARNING=1.
  95 EOF
  96         sleep 10
  97         printf "Proceeding with filter-branch...\n\n"
  98 fi
  99
 100 USAGE="[--setup <command>] [--subdirectory-filter <directory>] [--env-filter <command>]
 101         [--tree-filter <command>] [--index-filter <command>]
 102         [--parent-filter <command>] [--msg-filter <command>]
 103         [--commit-filter <command>] [--tag-name-filter <command>]
 104         [--original <namespace>]
 105         [-d <directory>] [-f | --force] [--state-branch <branch>]
 106         [--] [<rev-list options>...]"
 107
 108 OPTIONS_SPEC=
 109 . git-sh-setup
 110
 111 if [ "$(is_bare_repository)" = false ]; then
 112         require_clean_work_tree 'rewrite branches'
 113 fi
 114
 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 :
 131 do
 132         case "$1" in
 133         --)
 134                 shift
 135                 break
 136                 ;;
 137         --force|-f)
 138                 shift
 139                 force=t
 140                 continue
 141                 ;;
 142         --remap-to-ancestor)
 143                 # deprecated ($remap_to_ancestor is set now automatically)
 144                 shift
 145                 remap_to_ancestor=t
 146                 continue
 147                 ;;
 148         --prune-empty)
 149                 shift
 150                 prune_empty=t
 151                 continue
 152                 ;;
 153         -*)
 154                 ;;
 155         *)
 156                 break;
 157         esac
 158
 159         # all switches take one argument
 160         ARG="$1"
 161         case "$#" in 1) usage ;; esac
 162         shift
 163         OPTARG="$1"
 164         shift
 165
 166         case "$ARG" in
 167         -d)
 168                 tempdir="$OPTARG"
 169                 ;;
 170         --setup)
 171                 filter_setup="$OPTARG"
 172                 ;;
 173         --subdirectory-filter)
 174                 filter_subdir="$OPTARG"
 175                 remap_to_ancestor=t
 176                 ;;
 177         --env-filter)
 178                 filter_env="$OPTARG"
 179                 ;;
 180         --tree-filter)
 181                 filter_tree="$OPTARG"
 182                 ;;
 183         --index-filter)
 184                 filter_index="$OPTARG"
 185                 ;;
 186         --parent-filter)
 187                 filter_parent="$OPTARG"
 188                 ;;
 189         --msg-filter)
 190                 filter_msg="$OPTARG"
 191                 ;;
 192         --commit-filter)
 193                 filter_commit="$functions; $OPTARG"
 194                 ;;
 195         --tag-name-filter)
 196                 filter_tag_name="$OPTARG"
 197                 ;;
 198         --original)
 199                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 200                 ;;
 201         --state-branch)
 202                 state_branch="$OPTARG"
 203                 ;;
 204         *)
 205                 usage
 206                 ;;
 207         esac
 208 done
 209
 210 case "$prune_empty,$filter_commit" in
 211 ,)
 212         filter_commit='git commit-tree "$@"';;
 213 t,)
 214         filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
 215 ,*)
 216         ;;
 217 *)
 218         die "Cannot set --prune-empty and --commit-filter at the same time"
 219 esac
 220
 221 case "$force" in
 222 t)
 223         rm -rf "$tempdir"
 224 ;;
 225 '')
 226         test -d "$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="$(pwd)" ||
 234 die ""
 235
 236 # Remove tempdir on exit
 237 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
 238
 239 ORIG_GIT_DIR="$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"
 248
 249 GIT_WORK_TREE=.
 250 export GIT_DIR GIT_WORK_TREE
 251
 252 # Make sure refs/original is empty
 253 git for-each-ref > "$tempdir"/backup-refs || exit
 254 while read sha1 type name
 255 do
 256         case "$force,$name" in
 257         ,$orig_namespace*)
 258                 die "Cannot create a new backup.
 259 A previous backup already exists in $orig_namespace
 260 Force overwriting the backup with -f"
 261         ;;
 262         t,$orig_namespace*)
 263                 git update-ref -d "$name" $sha1
 264         ;;
 265         esac
 266 done < "$tempdir"/backup-refs
 267
 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
 272 do
 273         case "$ref" in ^?*) continue ;; esac
 274
 275         if git rev-parse --verify "$ref"^0 >/dev/null 2>&1
 276         then
 277                 echo "$ref"
 278         else
 279                 warn "WARNING: not rewriting '$ref' (not a committish)"
 280         fi
 281 done >"$tempdir"/heads <"$tempdir"/raw-refs
 282
 283 test -s "$tempdir"/heads ||
 284         die "You must specify a ref to rewrite."
 285
 286 GIT_INDEX_FILE="$(pwd)/../index"
 287 export GIT_INDEX_FILE
 288
 289 # map old->new commit ids for rewriting parents
 290 mkdir ../map || die "Could not create map/ directory"
 291
 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
 299                 git show "$state_commit:filter.map" >"$tempdir"/filter-map ||
 300                         die "Unable to load state from $state_branch:filter.map"
 301                 while read line
 302                 do
 303                         case "$line" in
 304                         *:*)
 305                                 echo "${line%:*}" >../map/"${line#*:}";;
 306                         *)
 307                                 die "Unable to load state from $state_branch:filter.map";;
 308                         esac
 309                 done <"$tempdir"/filter-map
 310         else
 311                 echo "Branch $state_branch does not exist. Will create" 1>&2
 312         fi
 313 fi
 314
 315 # we need "--" only if there are no path arguments in $@
 316 nonrevs=$(git rev-parse --no-revs "$@") || exit
 317 if test -z "$nonrevs"
 318 then
 319         dashdash=--
 320 else
 321         dashdash=
 322         remap_to_ancestor=t
 323 fi
 324
 325 git rev-parse --revs-only "$@" >../parse
 326
 327 case "$filter_subdir" in
 328 "")
 329         eval set -- "$(git rev-parse --sq --no-revs "$@")"
 330         ;;
 331 *)
 332         eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
 333                 "$filter_subdir")"
 334         ;;
 335 esac
 336
 337 git rev-list --reverse --topo-order --default HEAD \
 338         --parents --simplify-merges --stdin "$@" <../parse >../revs ||
 339         die "Could not get the commits"
 340 commits=$(wc -l <../revs | tr -d " ")
 341
 342 test $commits -eq 0 && die_with_status 2 "Found nothing to rewrite"
 343
 344 # Rewrite the commits
 345 report_progress ()
 346 {
 347         if test -n "$progress" &&
 348                 test $git_filter_branch__commit_count -gt $next_sample_at
 349         then
 350                 count=$git_filter_branch__commit_count
 351
 352                 now=$(date +%s)
 353                 elapsed=$(($now - $start_timestamp))
 354                 remaining=$(( ($commits - $count) * $elapsed / $count ))
 355                 if test $elapsed -gt 0
 356                 then
 357                         next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
 358                 else
 359                         next_sample_at=$(($next_sample_at + 1))
 360                 fi
 361                 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
 362         fi
 363         printf "\rRewrite $commit ($count/$commits)$progress    "
 364 }
 365
 366 git_filter_branch__commit_count=0
 367
 368 progress= start_timestamp=
 369 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
 370 then
 371         next_sample_at=0
 372         progress="dummy to ensure this is not empty"
 373         start_timestamp=$(date '+%s')
 374 fi
 375
 376 if test -n "$filter_index" ||
 377    test -n "$filter_tree" ||
 378    test -n "$filter_subdir"
 379 then
 380         need_index=t
 381 else
 382         need_index=
 383 fi
 384
 385 eval "$filter_setup" < /dev/null ||
 386         die "filter setup failed: $filter_setup"
 387
 388 while read commit parents; do
 389         git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
 390
 391         report_progress
 392         test -f "$workdir"/../map/$commit && continue
 393
 394         case "$filter_subdir" in
 395         "")
 396                 if test -n "$need_index"
 397                 then
 398                         GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
 399                 fi
 400                 ;;
 401         *)
 402                 # The commit may not have the subdirectory at all
 403                 err=$(GIT_ALLOW_NULL_SHA1=1 \
 404                       git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 405                         if ! git rev-parse -q --verify $commit:"$filter_subdir"
 406                         then
 407                                 rm -f "$GIT_INDEX_FILE"
 408                         else
 409                                 echo >&2 "$err"
 410                                 false
 411                         fi
 412                 }
 413         esac || die "Could not initialize the index"
 414
 415         GIT_COMMIT=$commit
 416         export GIT_COMMIT
 417         git cat-file commit "$commit" >../commit ||
 418                 die "Cannot read commit $commit"
 419
 420         eval "$(set_ident <../commit)" ||
 421                 die "setting author/committer failed for commit $commit"
 422         eval "$filter_env" < /dev/null ||
 423                 die "env filter failed: $filter_env"
 424
 425         if [ "$filter_tree" ]; then
 426                 git checkout-index -f -u -a ||
 427                         die "Could not checkout the index"
 428                 # files that $commit removed are now still in the working tree;
 429                 # remove them, else they would be added again
 430                 git clean -d -q -f -x
 431                 eval "$filter_tree" < /dev/null ||
 432                         die "tree filter failed: $filter_tree"
 433
 434                 (
 435                         git diff-index -r --name-only --ignore-submodules $commit -- &&
 436                         git ls-files --others
 437                 ) > "$tempdir"/tree-state || exit
 438                 git update-index --add --replace --remove --stdin \
 439                         < "$tempdir"/tree-state || exit
 440         fi
 441
 442         eval "$filter_index" < /dev/null ||
 443                 die "index filter failed: $filter_index"
 444
 445         parentstr=
 446         for parent in $parents; do
 447                 for reparent in $(map "$parent"); do
 448                         case "$parentstr " in
 449                         *" -p $reparent "*)
 450                                 ;;
 451                         *)
 452                                 parentstr="$parentstr -p $reparent"
 453                                 ;;
 454                         esac
 455                 done
 456         done
 457         if [ "$filter_parent" ]; then
 458                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 459                                 die "parent filter failed: $filter_parent"
 460         fi
 461
 462         {
 463                 while IFS='' read -r header_line && test -n "$header_line"
 464                 do
 465                         # skip header lines...
 466                         :;
 467                 done
 468                 # and output the actual commit message
 469                 cat
 470         } <../commit |
 471                 eval "$filter_msg" > ../message ||
 472                         die "msg filter failed: $filter_msg"
 473
 474         if test -n "$need_index"
 475         then
 476                 tree=$(git write-tree)
 477         else
 478                 tree=$(git rev-parse "$commit^{tree}")
 479         fi
 480         workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
 481                 "$tree" $parentstr < ../message > ../map/$commit ||
 482                         die "could not write rewritten commit"
 483 done <../revs
 484
 485 # If we are filtering for paths, as in the case of a subdirectory
 486 # filter, it is possible that a specified head is not in the set of
 487 # rewritten commits, because it was pruned by the revision walker.
 488 # Ancestor remapping fixes this by mapping these heads to the unique
 489 # nearest ancestor that survived the pruning.
 490
 491 if test "$remap_to_ancestor" = t
 492 then
 493         while read ref
 494         do
 495                 sha1=$(git rev-parse "$ref"^0)
 496                 test -f "$workdir"/../map/$sha1 && continue
 497                 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
 498                 test "$ancestor" && echo $(map $ancestor) >"$workdir"/../map/$sha1
 499         done < "$tempdir"/heads
 500 fi
 501
 502 # Finally update the refs
 503
 504 echo
 505 while read ref
 506 do
 507         # avoid rewriting a ref twice
 508         test -f "$orig_namespace$ref" && continue
 509
 510         sha1=$(git rev-parse "$ref"^0)
 511         rewritten=$(map $sha1)
 512
 513         test $sha1 = "$rewritten" &&
 514                 warn "WARNING: Ref '$ref' is unchanged" &&
 515                 continue
 516
 517         case "$rewritten" in
 518         '')
 519                 echo "Ref '$ref' was deleted"
 520                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 521                         die "Could not delete $ref"
 522         ;;
 523         *)
 524                 echo "Ref '$ref' was rewritten"
 525                 if ! git update-ref -m "filter-branch: rewrite" \
 526                                         "$ref" $rewritten $sha1 2>/dev/null; then
 527                         if test $(git cat-file -t "$ref") = tag; then
 528                                 if test -z "$filter_tag_name"; then
 529                                         warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
 530                                         warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
 531                                 fi
 532                         else
 533                                 die "Could not rewrite $ref"
 534                         fi
 535                 fi
 536         ;;
 537         esac
 538         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
 539                  exit
 540 done < "$tempdir"/heads
 541
 542 # TODO: This should possibly go, with the semantics that all positive given
 543 #       refs are updated, and their original heads stored in refs/original/
 544 # Filter tags
 545
 546 if [ "$filter_tag_name" ]; then
 547         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 548         while read sha1 type ref; do
 549                 ref="${ref#refs/tags/}"
 550                 # XXX: Rewrite tagged trees as well?
 551                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 552                         continue;
 553                 fi
 554
 555                 if [ "$type" = "tag" ]; then
 556                         # Dereference to a commit
 557                         sha1t="$sha1"
 558                         sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
 559                 fi
 560
 561                 [ -f "../map/$sha1" ] || continue
 562                 new_sha1="$(cat "../map/$sha1")"
 563                 GIT_COMMIT="$sha1"
 564                 export GIT_COMMIT
 565                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 566                         die "tag name filter failed: $filter_tag_name"
 567
 568                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 569
 570                 if [ "$type" = "tag" ]; then
 571                         new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
 572                                                 "$new_sha1" "$new_ref"
 573                                 git cat-file tag "$ref" |
 574                                 sed -n \
 575                                     -e '1,/^$/{
 576                                           /^object /d
 577                                           /^type /d
 578                                           /^tag /d
 579                                         }' \
 580                                     -e '/^-----BEGIN PGP SIGNATURE-----/q' \
 581                                     -e 'p' ) |
 582                                 git hash-object -t tag -w --stdin) ||
 583                                 die "Could not create new tag object for $ref"
 584                         if git cat-file tag "$ref" | \
 585                            grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
 586                         then
 587                                 warn "gpg signature stripped from tag object $sha1t"
 588                         fi
 589                 fi
 590
 591                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 592                         die "Could not write tag $new_ref"
 593         done
 594 fi
 595
 596 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 597 unset GIT_AUTHOR_NAME GIT_AUTHOR_EMAIL GIT_AUTHOR_DATE
 598 unset GIT_COMMITTER_NAME GIT_COMMITTER_EMAIL GIT_COMMITTER_DATE
 599 test -z "$ORIG_GIT_DIR" || {
 600         GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 601 }
 602 test -z "$ORIG_GIT_WORK_TREE" || {
 603         GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 604         export GIT_WORK_TREE
 605 }
 606 test -z "$ORIG_GIT_INDEX_FILE" || {
 607         GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 608         export GIT_INDEX_FILE
 609 }
 610 test -z "$ORIG_GIT_AUTHOR_NAME" || {
 611         GIT_AUTHOR_NAME="$ORIG_GIT_AUTHOR_NAME" &&
 612         export GIT_AUTHOR_NAME
 613 }
 614 test -z "$ORIG_GIT_AUTHOR_EMAIL" || {
 615         GIT_AUTHOR_EMAIL="$ORIG_GIT_AUTHOR_EMAIL" &&
 616         export GIT_AUTHOR_EMAIL
 617 }
 618 test -z "$ORIG_GIT_AUTHOR_DATE" || {
 619         GIT_AUTHOR_DATE="$ORIG_GIT_AUTHOR_DATE" &&
 620         export GIT_AUTHOR_DATE
 621 }
 622 test -z "$ORIG_GIT_COMMITTER_NAME" || {
 623         GIT_COMMITTER_NAME="$ORIG_GIT_COMMITTER_NAME" &&
 624         export GIT_COMMITTER_NAME
 625 }
 626 test -z "$ORIG_GIT_COMMITTER_EMAIL" || {
 627         GIT_COMMITTER_EMAIL="$ORIG_GIT_COMMITTER_EMAIL" &&
 628         export GIT_COMMITTER_EMAIL
 629 }
 630 test -z "$ORIG_GIT_COMMITTER_DATE" || {
 631         GIT_COMMITTER_DATE="$ORIG_GIT_COMMITTER_DATE" &&
 632         export GIT_COMMITTER_DATE
 633 }
 634
 635 if test -n "$state_branch"
 636 then
 637         echo "Saving rewrite state to $state_branch" 1>&2
 638         state_blob=$(
 639                 for file in ../map/*
 640                 do
 641                         from_commit=$(basename "$file")
 642                         to_commit=$(cat "$file")
 643                         echo "$from_commit:$to_commit"
 644                 done | git hash-object -w --stdin || die "Unable to save state"
 645         )
 646         state_tree=$(printf '100644 blob %s\tfilter.map\n' "$state_blob" | git mktree)
 647         if test -n "$state_commit"
 648         then
 649                 state_commit=$(echo "Sync" | git commit-tree "$state_tree" -p "$state_commit")
 650         else
 651                 state_commit=$(echo "Sync" | git commit-tree "$state_tree" )
 652         fi
 653         git update-ref "$state_branch" "$state_commit"
 654 fi
 655
 656 cd "$orig_dir"
 657 rm -rf "$tempdir"
 658
 659 trap - 0
 660
 661 if [ "$(is_bare_repository)" = false ]; then
 662         git read-tree -u -m HEAD || exit
 663 fi
 664
 665 exit 0