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