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