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 "$1";
  39                 shift;
  40         done;
  41 }
  42
  43 # override die(): this version puts in an extra line break, so that
  44 # the progress is still visible
  45
  46 die()
  47 {
  48         echo >&2
  49         echo "$*" >&2
  50         exit 1
  51 }
  52 EOF
  53 )
  54
  55 eval "$functions"
  56
  57 # When piped a commit, output a script to set the ident of either
  58 # "author" or "committer
  59
  60 set_ident () {
  61         lid="$(echo "$1" | tr "[A-Z]" "[a-z]")"
  62         uid="$(echo "$1" | tr "[a-z]" "[A-Z]")"
  63         pick_id_script='
  64                 /^'$lid' /{
  65                         s/'\''/'\''\\'\'\''/g
  66                         h
  67                         s/^'$lid' \([^<]*\) <[^>]*> .*$/\1/
  68                         s/'\''/'\''\'\'\''/g
  69                         s/.*/GIT_'$uid'_NAME='\''&'\''; export GIT_'$uid'_NAME/p
  70
  71                         g
  72                         s/^'$lid' [^<]* <\([^>]*\)> .*$/\1/
  73                         s/'\''/'\''\'\'\''/g
  74                         s/.*/GIT_'$uid'_EMAIL='\''&'\''; export GIT_'$uid'_EMAIL/p
  75
  76                         g
  77                         s/^'$lid' [^<]* <[^>]*> \(.*\)$/\1/
  78                         s/'\''/'\''\'\'\''/g
  79                         s/.*/GIT_'$uid'_DATE='\''&'\''; export GIT_'$uid'_DATE/p
  80
  81                         q
  82                 }
  83         '
  84
  85         LANG=C LC_ALL=C sed -ne "$pick_id_script"
  86         # Ensure non-empty id name.
  87         echo "case \"\$GIT_${uid}_NAME\" in \"\") GIT_${uid}_NAME=\"\${GIT_${uid}_EMAIL%%@*}\" && export GIT_${uid}_NAME;; esac"
  88 }
  89
  90 USAGE="[--env-filter <command>] [--tree-filter <command>] \
  91 [--index-filter <command>] [--parent-filter <command>] \
  92 [--msg-filter <command>] [--commit-filter <command>] \
  93 [--tag-name-filter <command>] [--subdirectory-filter <directory>] \
  94 [--original <namespace>] [-d <directory>] [-f | --force] \
  95 [<rev-list options>...]"
  96
  97 OPTIONS_SPEC=
  98 . git-sh-setup
  99
 100 git diff-files --quiet &&
 101         git diff-index --cached --quiet HEAD -- ||
 102         die "Cannot rewrite branch(es) with a dirty working directory."
 103
 104 tempdir=.git-rewrite
 105 filter_env=
 106 filter_tree=
 107 filter_index=
 108 filter_parent=
 109 filter_msg=cat
 110 filter_commit='git commit-tree "$@"'
 111 filter_tag_name=
 112 filter_subdir=
 113 orig_namespace=refs/original/
 114 force=
 115 while :
 116 do
 117         case "$1" in
 118         --)
 119                 shift
 120                 break
 121                 ;;
 122         --force|-f)
 123                 shift
 124                 force=t
 125                 continue
 126                 ;;
 127         -*)
 128                 ;;
 129         *)
 130                 break;
 131         esac
 132
 133         # all switches take one argument
 134         ARG="$1"
 135         case "$#" in 1) usage ;; esac
 136         shift
 137         OPTARG="$1"
 138         shift
 139
 140         case "$ARG" in
 141         -d)
 142                 tempdir="$OPTARG"
 143                 ;;
 144         --env-filter)
 145                 filter_env="$OPTARG"
 146                 ;;
 147         --tree-filter)
 148                 filter_tree="$OPTARG"
 149                 ;;
 150         --index-filter)
 151                 filter_index="$OPTARG"
 152                 ;;
 153         --parent-filter)
 154                 filter_parent="$OPTARG"
 155                 ;;
 156         --msg-filter)
 157                 filter_msg="$OPTARG"
 158                 ;;
 159         --commit-filter)
 160                 filter_commit="$functions; $OPTARG"
 161                 ;;
 162         --tag-name-filter)
 163                 filter_tag_name="$OPTARG"
 164                 ;;
 165         --subdirectory-filter)
 166                 filter_subdir="$OPTARG"
 167                 ;;
 168         --original)
 169                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 170                 ;;
 171         *)
 172                 usage
 173                 ;;
 174         esac
 175 done
 176
 177 case "$force" in
 178 t)
 179         rm -rf "$tempdir"
 180 ;;
 181 '')
 182         test -d "$tempdir" &&
 183                 die "$tempdir already exists, please remove it"
 184 esac
 185 mkdir -p "$tempdir/t" &&
 186 tempdir="$(cd "$tempdir"; pwd)" &&
 187 cd "$tempdir/t" &&
 188 workdir="$(pwd)" ||
 189 die ""
 190
 191 # Remove tempdir on exit
 192 trap 'cd ../..; rm -rf "$tempdir"' 0
 193
 194 # Make sure refs/original is empty
 195 git for-each-ref > "$tempdir"/backup-refs
 196 while read sha1 type name
 197 do
 198         case "$force,$name" in
 199         ,$orig_namespace*)
 200                 die "Namespace $orig_namespace not empty"
 201         ;;
 202         t,$orig_namespace*)
 203                 git update-ref -d "$name" $sha1
 204         ;;
 205         esac
 206 done < "$tempdir"/backup-refs
 207
 208 ORIG_GIT_DIR="$GIT_DIR"
 209 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 210 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 211 GIT_WORK_TREE=.
 212 export GIT_DIR GIT_WORK_TREE
 213
 214 # The refs should be updated if their heads were rewritten
 215 git rev-parse --no-flags --revs-only --symbolic-full-name --default HEAD "$@" |
 216 sed -e '/^^/d' >"$tempdir"/heads
 217
 218 test -s "$tempdir"/heads ||
 219         die "Which ref do you want to rewrite?"
 220
 221 GIT_INDEX_FILE="$(pwd)/../index"
 222 export GIT_INDEX_FILE
 223 git read-tree || die "Could not seed the index"
 224
 225 ret=0
 226
 227 # map old->new commit ids for rewriting parents
 228 mkdir ../map || die "Could not create map/ directory"
 229
 230 case "$filter_subdir" in
 231 "")
 232         git rev-list --reverse --topo-order --default HEAD \
 233                 --parents "$@"
 234         ;;
 235 *)
 236         git rev-list --reverse --topo-order --default HEAD \
 237                 --parents --full-history "$@" -- "$filter_subdir"
 238 esac > ../revs || die "Could not get the commits"
 239 commits=$(wc -l <../revs | tr -d " ")
 240
 241 test $commits -eq 0 && die "Found nothing to rewrite"
 242
 243 # Rewrite the commits
 244
 245 i=0
 246 while read commit parents; do
 247         i=$(($i+1))
 248         printf "\rRewrite $commit ($i/$commits)"
 249
 250         case "$filter_subdir" in
 251         "")
 252                 git read-tree -i -m $commit
 253                 ;;
 254         *)
 255                 # The commit may not have the subdirectory at all
 256                 err=$(git read-tree -i -m $commit:"$filter_subdir" 2>&1) || {
 257                         if ! git rev-parse --verify $commit:"$filter_subdir" 2>/dev/null
 258                         then
 259                                 rm -f "$GIT_INDEX_FILE"
 260                         else
 261                                 echo >&2 "$err"
 262                                 false
 263                         fi
 264                 }
 265         esac || die "Could not initialize the index"
 266
 267         GIT_COMMIT=$commit
 268         export GIT_COMMIT
 269         git cat-file commit "$commit" >../commit ||
 270                 die "Cannot read commit $commit"
 271
 272         eval "$(set_ident AUTHOR <../commit)" ||
 273                 die "setting author failed for commit $commit"
 274         eval "$(set_ident COMMITTER <../commit)" ||
 275                 die "setting committer failed for commit $commit"
 276         eval "$filter_env" < /dev/null ||
 277                 die "env filter failed: $filter_env"
 278
 279         if [ "$filter_tree" ]; then
 280                 git checkout-index -f -u -a ||
 281                         die "Could not checkout the index"
 282                 # files that $commit removed are now still in the working tree;
 283                 # remove them, else they would be added again
 284                 git ls-files -z --others | xargs -0 rm -f
 285                 eval "$filter_tree" < /dev/null ||
 286                         die "tree filter failed: $filter_tree"
 287
 288                 (
 289                         git diff-index -r --name-only $commit
 290                         git ls-files --others
 291                 ) |
 292                 git update-index --add --replace --remove --stdin
 293         fi
 294
 295         eval "$filter_index" < /dev/null ||
 296                 die "index filter failed: $filter_index"
 297
 298         parentstr=
 299         for parent in $parents; do
 300                 for reparent in $(map "$parent"); do
 301                         parentstr="$parentstr -p $reparent"
 302                 done
 303         done
 304         if [ "$filter_parent" ]; then
 305                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 306                                 die "parent filter failed: $filter_parent"
 307         fi
 308
 309         sed -e '1,/^$/d' <../commit | \
 310                 eval "$filter_msg" > ../message ||
 311                         die "msg filter failed: $filter_msg"
 312         sh -c "$filter_commit" "git commit-tree" \
 313                 $(git write-tree) $parentstr < ../message > ../map/$commit
 314 done <../revs
 315
 316 # In case of a subdirectory filter, it is possible that a specified head
 317 # is not in the set of rewritten commits, because it was pruned by the
 318 # revision walker.  Fix it by mapping these heads to the next rewritten
 319 # ancestor(s), i.e. the boundaries in the set of rewritten commits.
 320
 321 # NEEDSWORK: we should sort the unmapped refs topologically first
 322 while read ref
 323 do
 324         sha1=$(git rev-parse "$ref"^0)
 325         test -f "$workdir"/../map/$sha1 && continue
 326         # Assign the boundarie(s) in the set of rewritten commits
 327         # as the replacement commit(s).
 328         # (This would look a bit nicer if --not --stdin worked.)
 329         for p in $( (cd "$workdir"/../map; ls | sed "s/^/^/") |
 330                 git rev-list $ref --boundary --stdin |
 331                 sed -n "s/^-//p")
 332         do
 333                 map $p >> "$workdir"/../map/$sha1
 334         done
 335 done < "$tempdir"/heads
 336
 337 # Finally update the refs
 338
 339 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 340 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 341 echo
 342 while read ref
 343 do
 344         # avoid rewriting a ref twice
 345         test -f "$orig_namespace$ref" && continue
 346
 347         sha1=$(git rev-parse "$ref"^0)
 348         rewritten=$(map $sha1)
 349
 350         test $sha1 = "$rewritten" &&
 351                 warn "WARNING: Ref '$ref' is unchanged" &&
 352                 continue
 353
 354         case "$rewritten" in
 355         '')
 356                 echo "Ref '$ref' was deleted"
 357                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 358                         die "Could not delete $ref"
 359         ;;
 360         $_x40)
 361                 echo "Ref '$ref' was rewritten"
 362                 git update-ref -m "filter-branch: rewrite" \
 363                                 "$ref" $rewritten $sha1 ||
 364                         die "Could not rewrite $ref"
 365         ;;
 366         *)
 367                 # NEEDSWORK: possibly add -Werror, making this an error
 368                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 369                 warn "$rewritten"
 370                 warn "WARNING: Ref '$ref' points to the first one now."
 371                 rewritten=$(echo "$rewritten" | head -n 1)
 372                 git update-ref -m "filter-branch: rewrite to first" \
 373                                 "$ref" $rewritten $sha1 ||
 374                         die "Could not rewrite $ref"
 375         ;;
 376         esac
 377         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1
 378 done < "$tempdir"/heads
 379
 380 # TODO: This should possibly go, with the semantics that all positive given
 381 #       refs are updated, and their original heads stored in refs/original/
 382 # Filter tags
 383
 384 if [ "$filter_tag_name" ]; then
 385         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 386         while read sha1 type ref; do
 387                 ref="${ref#refs/tags/}"
 388                 # XXX: Rewrite tagged trees as well?
 389                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 390                         continue;
 391                 fi
 392
 393                 if [ "$type" = "tag" ]; then
 394                         # Dereference to a commit
 395                         sha1t="$sha1"
 396                         sha1="$(git rev-parse "$sha1"^{commit} 2>/dev/null)" || continue
 397                 fi
 398
 399                 [ -f "../map/$sha1" ] || continue
 400                 new_sha1="$(cat "../map/$sha1")"
 401                 GIT_COMMIT="$sha1"
 402                 export GIT_COMMIT
 403                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 404                         die "tag name filter failed: $filter_tag_name"
 405
 406                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 407
 408                 if [ "$type" = "tag" ]; then
 409                         # Warn that we are not rewriting the tag object itself.
 410                         warn "unreferencing tag object $sha1t"
 411                 fi
 412
 413                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 414                         die "Could not write tag $new_ref"
 415         done
 416 fi
 417
 418 cd ../..
 419 rm -rf "$tempdir"
 420
 421 trap - 0
 422
 423 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 424 test -z "$ORIG_GIT_DIR" || GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 425 test -z "$ORIG_GIT_WORK_TREE" || GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 426         export GIT_WORK_TREE
 427 test -z "$ORIG_GIT_INDEX_FILE" || GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 428         export GIT_INDEX_FILE
 429 git read-tree -u -m HEAD
 430
 431 exit $ret