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