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_$1_DATE"
  76 }
  77
  78 set_ident () {
  79         parse_ident_from_commit author AUTHOR committer COMMITTER
  80         finish_ident AUTHOR
  81         finish_ident COMMITTER
  82 }
  83
  84 USAGE="[--setup <command>] [--env-filter <command>]
  85         [--tree-filter <command>] [--index-filter <command>]
  86         [--parent-filter <command>] [--msg-filter <command>]
  87         [--commit-filter <command>] [--tag-name-filter <command>]
  88         [--subdirectory-filter <directory>] [--original <namespace>]
  89         [-d <directory>] [-f | --force]
  90         [--] [<rev-list options>...]"
  91
  92 OPTIONS_SPEC=
  93 . git-sh-setup
  94
  95 if [ "$(is_bare_repository)" = false ]; then
  96         require_clean_work_tree 'rewrite branches'
  97 fi
  98
  99 tempdir=.git-rewrite
 100 filter_setup=
 101 filter_env=
 102 filter_tree=
 103 filter_index=
 104 filter_parent=
 105 filter_msg=cat
 106 filter_commit=
 107 filter_tag_name=
 108 filter_subdir=
 109 orig_namespace=refs/original/
 110 force=
 111 prune_empty=
 112 remap_to_ancestor=
 113 while :
 114 do
 115         case "$1" in
 116         --)
 117                 shift
 118                 break
 119                 ;;
 120         --force|-f)
 121                 shift
 122                 force=t
 123                 continue
 124                 ;;
 125         --remap-to-ancestor)
 126                 # deprecated ($remap_to_ancestor is set now automatically)
 127                 shift
 128                 remap_to_ancestor=t
 129                 continue
 130                 ;;
 131         --prune-empty)
 132                 shift
 133                 prune_empty=t
 134                 continue
 135                 ;;
 136         -*)
 137                 ;;
 138         *)
 139                 break;
 140         esac
 141
 142         # all switches take one argument
 143         ARG="$1"
 144         case "$#" in 1) usage ;; esac
 145         shift
 146         OPTARG="$1"
 147         shift
 148
 149         case "$ARG" in
 150         -d)
 151                 tempdir="$OPTARG"
 152                 ;;
 153         --setup)
 154                 filter_setup="$OPTARG"
 155                 ;;
 156         --env-filter)
 157                 filter_env="$OPTARG"
 158                 ;;
 159         --tree-filter)
 160                 filter_tree="$OPTARG"
 161                 ;;
 162         --index-filter)
 163                 filter_index="$OPTARG"
 164                 ;;
 165         --parent-filter)
 166                 filter_parent="$OPTARG"
 167                 ;;
 168         --msg-filter)
 169                 filter_msg="$OPTARG"
 170                 ;;
 171         --commit-filter)
 172                 filter_commit="$functions; $OPTARG"
 173                 ;;
 174         --tag-name-filter)
 175                 filter_tag_name="$OPTARG"
 176                 ;;
 177         --subdirectory-filter)
 178                 filter_subdir="$OPTARG"
 179                 remap_to_ancestor=t
 180                 ;;
 181         --original)
 182                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 183                 ;;
 184         *)
 185                 usage
 186                 ;;
 187         esac
 188 done
 189
 190 case "$prune_empty,$filter_commit" in
 191 ,)
 192         filter_commit='git commit-tree "$@"';;
 193 t,)
 194         filter_commit="$functions;"' git_commit_non_empty_tree "$@"';;
 195 ,*)
 196         ;;
 197 *)
 198         die "Cannot set --prune-empty and --commit-filter at the same time"
 199 esac
 200
 201 case "$force" in
 202 t)
 203         rm -rf "$tempdir"
 204 ;;
 205 '')
 206         test -d "$tempdir" &&
 207                 die "$tempdir already exists, please remove it"
 208 esac
 209 orig_dir=$(pwd)
 210 mkdir -p "$tempdir/t" &&
 211 tempdir="$(cd "$tempdir"; pwd)" &&
 212 cd "$tempdir/t" &&
 213 workdir="$(pwd)" ||
 214 die ""
 215
 216 # Remove tempdir on exit
 217 trap 'cd "$orig_dir"; rm -rf "$tempdir"' 0
 218
 219 ORIG_GIT_DIR="$GIT_DIR"
 220 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 221 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 222 GIT_WORK_TREE=.
 223 export GIT_DIR GIT_WORK_TREE
 224
 225 # Make sure refs/original is empty
 226 git for-each-ref > "$tempdir"/backup-refs || exit
 227 while read sha1 type name
 228 do
 229         case "$force,$name" in
 230         ,$orig_namespace*)
 231                 die "Cannot create a new backup.
 232 A previous backup already exists in $orig_namespace
 233 Force overwriting the backup with -f"
 234         ;;
 235         t,$orig_namespace*)
 236                 git update-ref -d "$name" $sha1
 237         ;;
 238         esac
 239 done < "$tempdir"/backup-refs
 240
 241 # The refs should be updated if their heads were rewritten
 242 git rev-parse --no-flags --revs-only --symbolic-full-name \
 243         --default HEAD "$@" > "$tempdir"/raw-heads || exit
 244 sed -e '/^^/d' "$tempdir"/raw-heads >"$tempdir"/heads
 245
 246 test -s "$tempdir"/heads ||
 247         die "You must specify a ref to rewrite."
 248
 249 GIT_INDEX_FILE="$(pwd)/../index"
 250 export GIT_INDEX_FILE
 251
 252 # map old->new commit ids for rewriting parents
 253 mkdir ../map || die "Could not create map/ directory"
 254
 255 # we need "--" only if there are no path arguments in $@
 256 nonrevs=$(git rev-parse --no-revs "$@") || exit
 257 if test -z "$nonrevs"
 258 then
 259         dashdash=--
 260 else
 261         dashdash=
 262         remap_to_ancestor=t
 263 fi
 264
 265 git rev-parse --revs-only "$@" >../parse
 266
 267 case "$filter_subdir" in
 268 "")
 269         eval set -- "$(git rev-parse --sq --no-revs "$@")"
 270         ;;
 271 *)
 272         eval set -- "$(git rev-parse --sq --no-revs "$@" $dashdash \
 273                 "$filter_subdir")"
 274         ;;
 275 esac
 276
 277 git rev-list --reverse --topo-order --default HEAD \
 278         --parents --simplify-merges --stdin "$@" <../parse >../revs ||
 279         die "Could not get the commits"
 280 commits=$(wc -l <../revs | tr -d " ")
 281
 282 test $commits -eq 0 && die "Found nothing to rewrite"
 283
 284 # Rewrite the commits
 285 report_progress ()
 286 {
 287         if test -n "$progress" &&
 288                 test $git_filter_branch__commit_count -gt $next_sample_at
 289         then
 290                 count=$git_filter_branch__commit_count
 291
 292                 now=$(date +%s)
 293                 elapsed=$(($now - $start_timestamp))
 294                 remaining=$(( ($commits - $count) * $elapsed / $count ))
 295                 if test $elapsed -gt 0
 296                 then
 297                         next_sample_at=$(( ($elapsed + 1) * $count / $elapsed ))
 298                 else
 299                         next_sample_at=$(($next_sample_at + 1))
 300                 fi
 301                 progress=" ($elapsed seconds passed, remaining $remaining predicted)"
 302         fi
 303         printf "\rRewrite $commit ($count/$commits)$progress    "
 304 }
 305
 306 git_filter_branch__commit_count=0
 307
 308 progress= start_timestamp=
 309 if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
 310 then
 311         next_sample_at=0
 312         progress="dummy to ensure this is not empty"
 313         start_timestamp=$(date '+%s')
 314 fi
 315
 316 if test -n "$filter_index" ||
 317    test -n "$filter_tree" ||
 318    test -n "$filter_subdir"
 319 then
 320         need_index=t
 321 else
 322         need_index=
 323 fi
 324
 325 eval "$filter_setup" < /dev/null ||
 326         die "filter setup failed: $filter_setup"
 327
 328 while read commit parents; do
 329         git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
 330
 331         report_progress
 332
 333         case "$filter_subdir" in
 334         "")
 335                 if test -n "$need_index"
 336                 then
 337                         GIT_ALLOW_NULL_SHA1=1 git read-tree -i -m $commit
 338                 fi
 339                 ;;
 340         *)
 341                 # The commit may not have the subdirectory at all
 342                 err=$(GIT_ALLOW_NULL_SHA1=1 \
 343                       git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 344                         if ! git rev-parse -q --verify $commit:"$filter_subdir"
 345                         then
 346                                 rm -f "$GIT_INDEX_FILE"
 347                         else
 348                                 echo >&2 "$err"
 349                                 false
 350                         fi
 351                 }
 352         esac || die "Could not initialize the index"
 353
 354         GIT_COMMIT=$commit
 355         export GIT_COMMIT
 356         git cat-file commit "$commit" >../commit ||
 357                 die "Cannot read commit $commit"
 358
 359         eval "$(set_ident <../commit)" ||
 360                 die "setting author/committer failed for commit $commit"
 361         eval "$filter_env" < /dev/null ||
 362                 die "env filter failed: $filter_env"
 363
 364         if [ "$filter_tree" ]; then
 365                 git checkout-index -f -u -a ||
 366                         die "Could not checkout the index"
 367                 # files that $commit removed are now still in the working tree;
 368                 # remove them, else they would be added again
 369                 git clean -d -q -f -x
 370                 eval "$filter_tree" < /dev/null ||
 371                         die "tree filter failed: $filter_tree"
 372
 373                 (
 374                         git diff-index -r --name-only --ignore-submodules $commit -- &&
 375                         git ls-files --others
 376                 ) > "$tempdir"/tree-state || exit
 377                 git update-index --add --replace --remove --stdin \
 378                         < "$tempdir"/tree-state || exit
 379         fi
 380
 381         eval "$filter_index" < /dev/null ||
 382                 die "index filter failed: $filter_index"
 383
 384         parentstr=
 385         for parent in $parents; do
 386                 for reparent in $(map "$parent"); do
 387                         case "$parentstr " in
 388                         *" -p $reparent "*)
 389                                 ;;
 390                         *)
 391                                 parentstr="$parentstr -p $reparent"
 392                                 ;;
 393                         esac
 394                 done
 395         done
 396         if [ "$filter_parent" ]; then
 397                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 398                                 die "parent filter failed: $filter_parent"
 399         fi
 400
 401         {
 402                 while IFS='' read -r header_line && test -n "$header_line"
 403                 do
 404                         # skip header lines...
 405                         :;
 406                 done
 407                 # and output the actual commit message
 408                 cat
 409         } <../commit |
 410                 eval "$filter_msg" > ../message ||
 411                         die "msg filter failed: $filter_msg"
 412
 413         if test -n "$need_index"
 414         then
 415                 tree=$(git write-tree)
 416         else
 417                 tree=$(git rev-parse "$commit^{tree}")
 418         fi
 419         workdir=$workdir @SHELL_PATH@ -c "$filter_commit" "git commit-tree" \
 420                 "$tree" $parentstr < ../message > ../map/$commit ||
 421                         die "could not write rewritten commit"
 422 done <../revs
 423
 424 # If we are filtering for paths, as in the case of a subdirectory
 425 # filter, it is possible that a specified head is not in the set of
 426 # rewritten commits, because it was pruned by the revision walker.
 427 # Ancestor remapping fixes this by mapping these heads to the unique
 428 # nearest ancestor that survived the pruning.
 429
 430 if test "$remap_to_ancestor" = t
 431 then
 432         while read ref
 433         do
 434                 sha1=$(git rev-parse "$ref"^0)
 435                 test -f "$workdir"/../map/$sha1 && continue
 436                 ancestor=$(git rev-list --simplify-merges -1 "$ref" "$@")
 437                 test "$ancestor" && echo $(map $ancestor) >> "$workdir"/../map/$sha1
 438         done < "$tempdir"/heads
 439 fi
 440
 441 # Finally update the refs
 442
 443 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 444 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 445 echo
 446 while read ref
 447 do
 448         # avoid rewriting a ref twice
 449         test -f "$orig_namespace$ref" && continue
 450
 451         sha1=$(git rev-parse "$ref"^0)
 452         rewritten=$(map $sha1)
 453
 454         test $sha1 = "$rewritten" &&
 455                 warn "WARNING: Ref '$ref' is unchanged" &&
 456                 continue
 457
 458         case "$rewritten" in
 459         '')
 460                 echo "Ref '$ref' was deleted"
 461                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 462                         die "Could not delete $ref"
 463         ;;
 464         $_x40)
 465                 echo "Ref '$ref' was rewritten"
 466                 if ! git update-ref -m "filter-branch: rewrite" \
 467                                         "$ref" $rewritten $sha1 2>/dev/null; then
 468                         if test $(git cat-file -t "$ref") = tag; then
 469                                 if test -z "$filter_tag_name"; then
 470                                         warn "WARNING: You said to rewrite tagged commits, but not the corresponding tag."
 471                                         warn "WARNING: Perhaps use '--tag-name-filter cat' to rewrite the tag."
 472                                 fi
 473                         else
 474                                 die "Could not rewrite $ref"
 475                         fi
 476                 fi
 477         ;;
 478         *)
 479                 # NEEDSWORK: possibly add -Werror, making this an error
 480                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 481                 warn "$rewritten"
 482                 warn "WARNING: Ref '$ref' points to the first one now."
 483                 rewritten=$(echo "$rewritten" | head -n 1)
 484                 git update-ref -m "filter-branch: rewrite to first" \
 485                                 "$ref" $rewritten $sha1 ||
 486                         die "Could not rewrite $ref"
 487         ;;
 488         esac
 489         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1 ||
 490                  exit
 491 done < "$tempdir"/heads
 492
 493 # TODO: This should possibly go, with the semantics that all positive given
 494 #       refs are updated, and their original heads stored in refs/original/
 495 # Filter tags
 496
 497 if [ "$filter_tag_name" ]; then
 498         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 499         while read sha1 type ref; do
 500                 ref="${ref#refs/tags/}"
 501                 # XXX: Rewrite tagged trees as well?
 502                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 503                         continue;
 504                 fi
 505
 506                 if [ "$type" = "tag" ]; then
 507                         # Dereference to a commit
 508                         sha1t="$sha1"
 509                         sha1="$(git rev-parse -q "$sha1"^{commit})" || continue
 510                 fi
 511
 512                 [ -f "../map/$sha1" ] || continue
 513                 new_sha1="$(cat "../map/$sha1")"
 514                 GIT_COMMIT="$sha1"
 515                 export GIT_COMMIT
 516                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 517                         die "tag name filter failed: $filter_tag_name"
 518
 519                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 520
 521                 if [ "$type" = "tag" ]; then
 522                         new_sha1=$( ( printf 'object %s\ntype commit\ntag %s\n' \
 523                                                 "$new_sha1" "$new_ref"
 524                                 git cat-file tag "$ref" |
 525                                 sed -n \
 526                                     -e '1,/^$/{
 527                                           /^object /d
 528                                           /^type /d
 529                                           /^tag /d
 530                                         }' \
 531                                     -e '/^-----BEGIN PGP SIGNATURE-----/q' \
 532                                     -e 'p' ) |
 533                                 git mktag) ||
 534                                 die "Could not create new tag object for $ref"
 535                         if git cat-file tag "$ref" | \
 536                            sane_grep '^-----BEGIN PGP SIGNATURE-----' >/dev/null 2>&1
 537                         then
 538                                 warn "gpg signature stripped from tag object $sha1t"
 539                         fi
 540                 fi
 541
 542                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 543                         die "Could not write tag $new_ref"
 544         done
 545 fi
 546
 547 cd "$orig_dir"
 548 rm -rf "$tempdir"
 549
 550 trap - 0
 551
 552 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 553 test -z "$ORIG_GIT_DIR" || {
 554         GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 555 }
 556 test -z "$ORIG_GIT_WORK_TREE" || {
 557         GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 558         export GIT_WORK_TREE
 559 }
 560 test -z "$ORIG_GIT_INDEX_FILE" || {
 561         GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 562         export GIT_INDEX_FILE
 563 }
 564
 565 if [ "$(is_bare_repository)" = false ]; then
 566         git read-tree -u -m HEAD || exit
 567 fi
 568
 569 exit 0