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         test $# = 0 && usage
 118         case "$1" in
 119         --)
 120                 shift
 121                 break
 122                 ;;
 123         --force|-f)
 124                 shift
 125                 force=t
 126                 continue
 127                 ;;
 128         -*)
 129                 ;;
 130         *)
 131                 break;
 132         esac
 133
 134         # all switches take one argument
 135         ARG="$1"
 136         case "$#" in 1) usage ;; esac
 137         shift
 138         OPTARG="$1"
 139         shift
 140
 141         case "$ARG" in
 142         -d)
 143                 tempdir="$OPTARG"
 144                 ;;
 145         --env-filter)
 146                 filter_env="$OPTARG"
 147                 ;;
 148         --tree-filter)
 149                 filter_tree="$OPTARG"
 150                 ;;
 151         --index-filter)
 152                 filter_index="$OPTARG"
 153                 ;;
 154         --parent-filter)
 155                 filter_parent="$OPTARG"
 156                 ;;
 157         --msg-filter)
 158                 filter_msg="$OPTARG"
 159                 ;;
 160         --commit-filter)
 161                 filter_commit="$functions; $OPTARG"
 162                 ;;
 163         --tag-name-filter)
 164                 filter_tag_name="$OPTARG"
 165                 ;;
 166         --subdirectory-filter)
 167                 filter_subdir="$OPTARG"
 168                 ;;
 169         --original)
 170                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 171                 ;;
 172         *)
 173                 usage
 174                 ;;
 175         esac
 176 done
 177
 178 case "$force" in
 179 t)
 180         rm -rf "$tempdir"
 181 ;;
 182 '')
 183         test -d "$tempdir" &&
 184                 die "$tempdir already exists, please remove it"
 185 esac
 186 mkdir -p "$tempdir/t" &&
 187 tempdir="$(cd "$tempdir"; pwd)" &&
 188 cd "$tempdir/t" &&
 189 workdir="$(pwd)" ||
 190 die ""
 191
 192 # Make sure refs/original is empty
 193 git for-each-ref > "$tempdir"/backup-refs
 194 while read sha1 type name
 195 do
 196         case "$force,$name" in
 197         ,$orig_namespace*)
 198                 die "Namespace $orig_namespace not empty"
 199         ;;
 200         t,$orig_namespace*)
 201                 git update-ref -d "$name" $sha1
 202         ;;
 203         esac
 204 done < "$tempdir"/backup-refs
 205
 206 ORIG_GIT_DIR="$GIT_DIR"
 207 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 208 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 209 GIT_WORK_TREE=.
 210 export GIT_DIR GIT_WORK_TREE
 211
 212 # The refs should be updated if their heads were rewritten
 213 git rev-parse --no-flags --revs-only --symbolic-full-name "$@" |
 214 sed -e '/^^/d' >"$tempdir"/heads
 215
 216 test -s "$tempdir"/heads ||
 217         die "Which ref do you want to rewrite?"
 218
 219 GIT_INDEX_FILE="$(pwd)/../index"
 220 export GIT_INDEX_FILE
 221 git read-tree || die "Could not seed the index"
 222
 223 ret=0
 224
 225 # map old->new commit ids for rewriting parents
 226 mkdir ../map || die "Could not create map/ directory"
 227
 228 case "$filter_subdir" in
 229 "")
 230         git rev-list --reverse --topo-order --default HEAD \
 231                 --parents "$@"
 232         ;;
 233 *)
 234         git rev-list --reverse --topo-order --default HEAD \
 235                 --parents --full-history "$@" -- "$filter_subdir"
 236 esac > ../revs || die "Could not get the commits"
 237 commits=$(wc -l <../revs | tr -d " ")
 238
 239 test $commits -eq 0 && die "Found nothing to rewrite"
 240
 241 # Rewrite the commits
 242
 243 i=0
 244 while read commit parents; do
 245         i=$(($i+1))
 246         printf "\rRewrite $commit ($i/$commits)"
 247
 248         case "$filter_subdir" in
 249         "")
 250                 git read-tree -i -m $commit
 251                 ;;
 252         *)
 253                 git read-tree -i -m $commit:"$filter_subdir"
 254         esac || die "Could not initialize the index"
 255
 256         GIT_COMMIT=$commit
 257         export GIT_COMMIT
 258         git cat-file commit "$commit" >../commit ||
 259                 die "Cannot read commit $commit"
 260
 261         eval "$(set_ident AUTHOR <../commit)" ||
 262                 die "setting author failed for commit $commit"
 263         eval "$(set_ident COMMITTER <../commit)" ||
 264                 die "setting committer failed for commit $commit"
 265         eval "$filter_env" < /dev/null ||
 266                 die "env filter failed: $filter_env"
 267
 268         if [ "$filter_tree" ]; then
 269                 git checkout-index -f -u -a ||
 270                         die "Could not checkout the index"
 271                 # files that $commit removed are now still in the working tree;
 272                 # remove them, else they would be added again
 273                 git ls-files -z --others | xargs -0 rm -f
 274                 eval "$filter_tree" < /dev/null ||
 275                         die "tree filter failed: $filter_tree"
 276
 277                 git diff-index -r $commit | cut -f 2- | tr '\012' '\000' | \
 278                         xargs -0 git update-index --add --replace --remove
 279                 git ls-files -z --others | \
 280                         xargs -0 git update-index --add --replace --remove
 281         fi
 282
 283         eval "$filter_index" < /dev/null ||
 284                 die "index filter failed: $filter_index"
 285
 286         parentstr=
 287         for parent in $parents; do
 288                 for reparent in $(map "$parent"); do
 289                         parentstr="$parentstr -p $reparent"
 290                 done
 291         done
 292         if [ "$filter_parent" ]; then
 293                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 294                                 die "parent filter failed: $filter_parent"
 295         fi
 296
 297         sed -e '1,/^$/d' <../commit | \
 298                 eval "$filter_msg" > ../message ||
 299                         die "msg filter failed: $filter_msg"
 300         sh -c "$filter_commit" "git commit-tree" \
 301                 $(git write-tree) $parentstr < ../message > ../map/$commit
 302 done <../revs
 303
 304 # In case of a subdirectory filter, it is possible that a specified head
 305 # is not in the set of rewritten commits, because it was pruned by the
 306 # revision walker.  Fix it by mapping these heads to the next rewritten
 307 # ancestor(s), i.e. the boundaries in the set of rewritten commits.
 308
 309 # NEEDSWORK: we should sort the unmapped refs topologically first
 310 while read ref
 311 do
 312         sha1=$(git rev-parse "$ref"^0)
 313         test -f "$workdir"/../map/$sha1 && continue
 314         # Assign the boundarie(s) in the set of rewritten commits
 315         # as the replacement commit(s).
 316         # (This would look a bit nicer if --not --stdin worked.)
 317         for p in $( (cd "$workdir"/../map; ls | sed "s/^/^/") |
 318                 git rev-list $ref --boundary --stdin |
 319                 sed -n "s/^-//p")
 320         do
 321                 map $p >> "$workdir"/../map/$sha1
 322         done
 323 done < "$tempdir"/heads
 324
 325 # Finally update the refs
 326
 327 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 328 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 329 echo
 330 while read ref
 331 do
 332         # avoid rewriting a ref twice
 333         test -f "$orig_namespace$ref" && continue
 334
 335         sha1=$(git rev-parse "$ref"^0)
 336         rewritten=$(map $sha1)
 337
 338         test $sha1 = "$rewritten" &&
 339                 warn "WARNING: Ref '$ref' is unchanged" &&
 340                 continue
 341
 342         case "$rewritten" in
 343         '')
 344                 echo "Ref '$ref' was deleted"
 345                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 346                         die "Could not delete $ref"
 347         ;;
 348         $_x40)
 349                 echo "Ref '$ref' was rewritten"
 350                 git update-ref -m "filter-branch: rewrite" \
 351                                 "$ref" $rewritten $sha1 ||
 352                         die "Could not rewrite $ref"
 353         ;;
 354         *)
 355                 # NEEDSWORK: possibly add -Werror, making this an error
 356                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 357                 warn "$rewritten"
 358                 warn "WARNING: Ref '$ref' points to the first one now."
 359                 rewritten=$(echo "$rewritten" | head -n 1)
 360                 git update-ref -m "filter-branch: rewrite to first" \
 361                                 "$ref" $rewritten $sha1 ||
 362                         die "Could not rewrite $ref"
 363         ;;
 364         esac
 365         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1
 366 done < "$tempdir"/heads
 367
 368 # TODO: This should possibly go, with the semantics that all positive given
 369 #       refs are updated, and their original heads stored in refs/original/
 370 # Filter tags
 371
 372 if [ "$filter_tag_name" ]; then
 373         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 374         while read sha1 type ref; do
 375                 ref="${ref#refs/tags/}"
 376                 # XXX: Rewrite tagged trees as well?
 377                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 378                         continue;
 379                 fi
 380
 381                 if [ "$type" = "tag" ]; then
 382                         # Dereference to a commit
 383                         sha1t="$sha1"
 384                         sha1="$(git rev-parse "$sha1"^{commit} 2>/dev/null)" || continue
 385                 fi
 386
 387                 [ -f "../map/$sha1" ] || continue
 388                 new_sha1="$(cat "../map/$sha1")"
 389                 GIT_COMMIT="$sha1"
 390                 export GIT_COMMIT
 391                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 392                         die "tag name filter failed: $filter_tag_name"
 393
 394                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 395
 396                 if [ "$type" = "tag" ]; then
 397                         # Warn that we are not rewriting the tag object itself.
 398                         warn "unreferencing tag object $sha1t"
 399                 fi
 400
 401                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 402                         die "Could not write tag $new_ref"
 403         done
 404 fi
 405
 406 cd ../..
 407 rm -rf "$tempdir"
 408
 409 unset GIT_DIR GIT_WORK_TREE GIT_INDEX_FILE
 410 test -z "$ORIG_GIT_DIR" || GIT_DIR="$ORIG_GIT_DIR" && export GIT_DIR
 411 test -z "$ORIG_GIT_WORK_TREE" || GIT_WORK_TREE="$ORIG_GIT_WORK_TREE" &&
 412         export GIT_WORK_TREE
 413 test -z "$ORIG_GIT_INDEX_FILE" || GIT_INDEX_FILE="$ORIG_GIT_INDEX_FILE" &&
 414         export GIT_INDEX_FILE
 415 git read-tree -u -m HEAD
 416
 417 exit $ret