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