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