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_$1_DATE"
  76 }
  77
  78 set_ident () {
  79         parse_ident_from_commit author AUTHOR committer COMMITTER
  80         finish_ident AUTHOR
  81         finish_ident COMMITTER
  82 }
  83
  84 USAGE="[--env-filter <command>] [--tree-filter <command>]
  85         [--index-filter <command>] [--parent-filter <command>]
  86         [--msg-filter <command>] [--commit-filter <command>]
  87         [--tag-name-filter <command>] [--subdirectory-filter <directory>]
  88         [--original <namespace>] [-d <directory>] [-f | --force]
  89         [<rev-list options>...]"
  90
  91 OPTIONS_SPEC=
  92 . git-sh-setup
  93
  94 if [ "$(is_bare_repository)" = false ]; then
  95         require_clean_work_tree 'rewrite branches'
  96 fi
  97
  98 tempdir=.git-rewrite
  99 filter_env=
 100 filter_tree=
 101 filter_index=
 102 filter_parent=
 103 filter_msg=cat
 104 filter_commit=
 105 filter_tag_name=
 106 filter_subdir=
 107 orig_namespace=refs/original/
 108 force=
 109 prune_empty=
 110 remap_to_ancestor=
 111 while :
 112 do
 113         case "$1" in
 114         --)
 115                 shift
 116                 break
 117                 ;;
 118         --force|-f)
 119                 shift
 120                 force=t
 121                 continue
 122                 ;;
 123         --remap-to-ancestor)
 124                 # deprecated ($remap_to_ancestor is set now automatically)
 125                 shift
 126                 remap_to_ancestor=t
 127                 continue
 128                 ;;
 129         --prune-empty)
 130                 shift
 131                 prune_empty=t
 132                 continue
 133                 ;;
 134         -*)
 135                 ;;
 136         *)
 137                 break;
 138         esac
 139
 140         # all switches take one argument
 141         ARG="$1"
 142         case "$#" in 1) usage ;; esac
 143         shift
 144         OPTARG="$1"
 145         shift
 146
 147         case "$ARG" in
 148         -d)
 149                 tempdir="$OPTARG"
 150                 ;;
 151         --env-filter)
 152                 filter_env="$OPTARG"
 153                 ;;
 154         --tree-filter)
 155                 filter_tree="$OPTARG"
 156                 ;;
 157         --index-filter)
 158                 filter_index="$OPTARG"
 159                 ;;
 160         --parent-filter)
 161                 filter_parent="$OPTARG"
 162                 ;;
 163         --msg-filter)
 164                 filter_msg="$OPTARG"
 165                 ;;
 166         --commit-filter)
 167                 filter_commit="$functions; $OPTARG"
 168                 ;;
 169         --tag-name-filter)
 170                 filter_tag_name="$OPTARG"
 171                 ;;
 172         --subdirectory-filter)
 173                 filter_subdir="$OPTARG"
 174                 remap_to_ancestor=t
 175                 ;;
 176         --original)
 177                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 178                 ;;
 179         *)
 180                 usage
 181                 ;;
 182         esac
 183 done
 184
 185 case "$prune_empty,$filter_commit" in
 186 ,)
 187         filter_commit='git commit-tree "$@"';;
 188 t,)
 189         filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
 190 ,*)
 191         ;;
 192 *)
 193         die "Cannot set --prune-empty and --commit-filter at the same time"
 194 esac
 195
 196 case "$force" in
 197 t)
 198         rm -rf "$tempdir"
 199 ;;
 200 '')
 201         test -d "$tempdir" &&
 202                 die "$tempdir already exists, please remove it"
 203 esac
 204 orig_dir=$(pwd)
 205 mkdir -p "$tempdir/t" &&
 206 tempdir="$(cd "$tempdir"; pwd)" &&
 207 cd "$tempdir/t" &&
 208 workdir="$(pwd)" ||
 209 die ""
 210
 211 # Remove tempdir on exit
 212 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
 213
 214 ORIG_GIT_DIR="$GIT_DIR"
 215 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 216 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 217 GIT_WORK_TREE=.
 218 export GIT_DIR GIT_WORK_TREE
 219
 220 # Make sure refs/original is empty
 221 git for-each-ref > "$tempdir"/backup-refs || exit
 222 while read sha1 type name
 223 do
 224         case "$force,$name" in
 225         ,$orig_namespace*)
 226                 die "Cannot create a new backup.
 227 A previous backup already exists in $orig_namespace
 228 Force overwriting the backup with -f"
 229         ;;
 230         t,$orig_namespace*)
 231                 git update-ref -d "$name" $sha1
 232         ;;
 233         esac
 234 done < "$tempdir"/backup-refs
 235
 236 # The refs should be updated if their heads were rewritten
 237 git rev-parse --no-flags --revs-only --symbolic-full-name \
 238         --default HEAD "$@" > "$tempdir"/raw-heads || exit
 239 sed -e '/^^/d' "$tempdir"/raw-heads >"$tempdir"/heads
 240
 241 test -s "$tempdir"/heads ||
 242         die "Which ref do you want to rewrite?"
 243
 244 GIT_INDEX_FILE="$(pwd)/../index"
 245 export GIT_INDEX_FILE
 246
 247 # map old->new commit ids for rewriting parents
 248 mkdir ../map || die "Could not create map/ directory"
 249
 250 # we need "--" only if there are no path arguments in $@
 251 nonrevs=$(git rev-parse --no-revs "$@") || exit
 252 if test -z "$nonrevs"
 253 then
 254         dashdash=--
 255 else
 256         dashdash=
 257         remap_to_ancestor=t
 258 fi
 259
 260 git rev-parse --revs-only "$@" >../parse
 261
 262 case "$filter_subdir" in
 263 "")
 264         eval set -- "$(git rev-parse --sq --no-revs "$@")"
 265         ;;
 266 *)
 267         eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
 268                 "$filter_subdir")"
 269         ;;
 270 esac
 271
 272 git rev-list --reverse --topo-order --default HEAD \
 273         --parents --simplify-merges --stdin "$@" <../parse >../revs ||
 274         die "Could not get the commits"
 275 commits=$(wc -l <../revs | tr -d " ")
 276
 277 test $commits -eq 0 && die "Found nothing to rewrite"
 278
 279 # Rewrite the commits
 280 report_progress ()
 281 {
 282         if test -n "$progress" &&
 283                 test $git_filter_branch__commit_count -gt $next_sample_at
 284         then
 285                 count=$git_filter_branch__commit_count
 286
 287                 now=$(date +%s)
 288                 elapsed=$(($now - $start_timestamp))
 289                 remaining=$(( ($commits - $count) * $elapsed / $count ))
 290                 if test $elapsed -gt 0
 291                 then
 292                         next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
 293                 else
 294                         next_sample_at=$(($next_sample_at + 1))
 295                 fi
 296                 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
 297         fi
 298         printf "\rRewrite $commit ($count/$commits)$progress    "
 299 }
 300
 301 git_filter_branch__commit_count=0
 302
 303 progress= start_timestamp=
 304 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
 305 then
 306         next_sample_at=0
 307         progress="dummy to ensure this is not empty"
 308         start_timestamp=$(date '+%s')
 309 fi
 310
 311 if test -n "$filter_index" ||
 312    test -n "$filter_tree" ||
 313    test -n "$filter_subdir"
 314 then
 315         need_index=t
 316 else
 317         need_index=
 318 fi
 319
 320 while read commit parents; do
 321         git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
 322
 323         report_progress
 324
 325         case "$filter_subdir" in
 326         "")
 327                 if test -n "$need_index"
 328                 then
 329                         GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
 330                 fi
 331                 ;;
 332         *)
 333                 # The commit may not have the subdirectory at all
 334                 err=$(GIT_ALLOW_NULL_SHA1=1 \
 335                       git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 336                         if ! git rev-parse -q --verify $commit:"$filter_subdir"
 337                         then
 338                                 rm -f "$GIT_INDEX_FILE"
 339                         else
 340                                 echo >&2 "$err"
 341                                 false
 342                         fi
 343                 }
 344         esac || die "Could not initialize the index"
 345
 346         GIT_COMMIT=$commit
 347         export GIT_COMMIT
 348         git cat-file commit "$commit" >../commit ||
 349                 die "Cannot read commit $commit"
 350
 351         eval "$(set_ident <../commit)" ||
 352                 die "setting author/committer failed for commit $commit"
 353         eval "$filter_env" < /dev/null ||
 354                 die "env filter failed: $filter_env"
 355
 356         if [ "$filter_tree" ]; then
 357                 git checkout-index -f -u -a ||
 358                         die "Could not checkout the index"
 359                 # files that $commit removed are now still in the working tree;
 360                 # remove them, else they would be added again
 361                 git clean -d -q -f -x
 362                 eval "$filter_tree" < /dev/null ||
 363                         die "tree filter failed: $filter_tree"
 364
 365                 (
 366                         git diff-index -r --name-only --ignore-submodules $commit -- &&
 367                         git ls-files --others
 368                 ) > "$tempdir"/tree-state || exit
 369                 git update-index --add --replace --remove --stdin \
 370                         < "$tempdir"/tree-state || exit
 371         fi
 372
 373         eval "$filter_index" < /dev/null ||
 374                 die "index filter failed: $filter_index"
 375
 376         parentstr=
 377         for parent in $parents; do
 378                 for reparent in $(map "$parent"); do
 379                         case "$parentstr " in
 380                         *" -p $reparent "*)
 381                                 ;;
 382                         *)
 383                                 parentstr="$parentstr -p $reparent"
 384                                 ;;
 385                         esac
 386                 done
 387         done
 388         if [ "$filter_parent" ]; then
 389                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 390                                 die "parent filter failed: $filter_parent"
 391         fi
 392
 393         {
 394                 while IFS='' read -r header_line && test -n "$header_line"
 395                 do
 396                         # skip header lines...
 397                         :;
 398                 done
 399                 # and output the actual commit message
 400                 cat
 401         } <../commit |
 402                 eval "$filter_msg" > ../message ||
 403                         die "msg filter failed: $filter_msg"
 404
 405         if test -n "$need_index"
 406         then
 407                 tree=$(git write-tree)
 408         else
 409                 tree=$(git rev-parse "$commit^{tree}")
 410         fi
 411         workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
 412                 "$tree" $parentstr < ../message > ../map/$commit ||
 413                         die "could not write rewritten commit"
 414 done <../revs
 415
 416 # If we are filtering for paths, as in the case of a subdirectory
 417 # filter, it is possible that a specified head is not in the set of
 418 # rewritten commits, because it was pruned by the revision walker.
 419 # Ancestor remapping fixes this by mapping these heads to the unique
 420 # nearest ancestor that survived the pruning.
 421
 422 if test "$remap_to_ancestor" = t
 423 then
 424         while read ref
 425         do
 426                 sha1=$(git rev-parse "$ref"^0)
 427                 test -f "$workdir"/../map/$sha1 && continue
 428                 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
 429                 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
 430         done < "$tempdir"/heads
 431 fi
 432
 433 # Finally update the refs
 434
 435 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 436 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 437 echo
 438 while read ref
 439 do
 440         # avoid rewriting a ref twice
 441         test -f "$orig_namespace$ref" && continue
 442
 443         sha1=$(git rev-parse "$ref"^0)
 444         rewritten=$(map $sha1)
 445
 446         test $sha1 = "$rewritten" &&
 447                 warn "WARNING: Ref '$ref' is unchanged" &&
 448                 continue
 449
 450         case "$rewritten" in
 451         '')
 452                 echo "Ref '$ref' was deleted"
 453                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 454                         die "Could not delete $ref"
 455         ;;
 456         $_x40)
 457                 echo "Ref '$ref' was rewritten"
 458                 if ! git update-ref -m "filter-branch: rewrite" \
 459                                         "$ref" $rewritten $sha1 2>/dev/null; then
 460                         if test $(git cat-file -t "$ref") = tag; then
 461                                 if test -z "$filter_tag_name"; then
 462                                         warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
 463                                         warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
 464                                 fi
 465                         else
 466                                 die "Could not rewrite $ref"
 467                         fi
 468                 fi
 469         ;;
 470         *)
 471                 # NEEDSWORK: possibly add -Werror, making this an error
 472                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 473                 warn "$rewritten"
 474                 warn "WARNING: Ref '$ref' points to the first one now."
 475                 rewritten=$(echo "$rewritten" | head -n 1)
 476                 git update-ref -m "filter-branch: rewrite to first" \
 477                                 "$ref" $rewritten $sha1 ||
 478                         die "Could not rewrite $ref"
 479         ;;
 480         esac
 481         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
 482                  exit
 483 done < "$tempdir"/heads
 484
 485 # TODO: This should possibly go, with the semantics that all positive given
 486 #       refs are updated, and their original heads stored in refs/original/
 487 # Filter tags
 488
 489 if [ "$filter_tag_name" ]; then
 490         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 491         while read sha1 type ref; do
 492                 ref="${ref#refs/tags/}"
 493                 # XXX: Rewrite tagged trees as well?
 494                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 495                         continue;
 496                 fi
 497
 498                 if [ "$type" = "tag" ]; then
 499                         # Dereference to a commit
 500                         sha1t="$sha1"
 501                         sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
 502                 fi
 503
 504                 [ -f "../map/$sha1" ] || continue
 505                 new_sha1="$(cat "../map/$sha1")"
 506                 GIT_COMMIT="$sha1"
 507                 export GIT_COMMIT
 508                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 509                         die "tag name filter failed: $filter_tag_name"
 510
 511                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 512
 513                 if [ "$type" = "tag" ]; then
 514                         new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
 515                                                 "$new_sha1" "$new_ref"
 516                                 git cat-file tag "$ref" |
 517                                 sed -n \
 518                                     -e '1,/^$/{
 519                                           /^object /d
 520                                           /^type /d
 521                                           /^tag /d
 522                                         }' \
 523                                     -e '/^-----BEGIN PGP SIGNATURE-----/q' \
 524                                     -e 'p' ) |
 525                                 git mktag) ||
 526                                 die "Could not create new tag object for $ref"
 527                         if git cat-file tag "$ref" | \
 528                            sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
 529                         then
 530                                 warn "gpg signature stripped from tag object $sha1t"
 531                         fi
 532                 fi
 533
 534                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 535                         die "Could not write tag $new_ref"
 536         done
 537 fi
 538
 539 cd "$orig_dir"
 540 rm -rf "$tempdir"
 541
 542 trap - 0
 543
 544 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 545 test -z "$ORIG_GIT_DIR" || {
 546         GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 547 }
 548 test -z "$ORIG_GIT_WORK_TREE" || {
 549         GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 550         export GIT_WORK_TREE
 551 }
 552 test -z "$ORIG_GIT_INDEX_FILE" || {
 553         GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 554         export GIT_INDEX_FILE
 555 }
 556
 557 if [ "$(is_bare_repository)" = false ]; then
 558         git read-tree -u -m HEAD || exit
 559 fi
 560
 561 exit 0