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 warn () {
  12         echo "$*" >&2
  13 }
  14
  15 map()
  16 {
  17         # if it was not rewritten, take the original
  18         if test -r "$workdir/../map/$1"
  19         then
  20                 cat "$workdir/../map/$1"
  21         else
  22                 echo "$1"
  23         fi
  24 }
  25
  26 # if you run 'skip_commit "$@"' in a commit filter, it will print
  27 # the (mapped) parents, effectively skipping the commit.
  28
  29 skip_commit()
  30 {
  31         shift;
  32         while [ -n "$1" ];
  33         do
  34                 shift;
  35                 map "$1";
  36                 shift;
  37         done;
  38 }
  39
  40 # override die(): this version puts in an extra line break, so that
  41 # the progress is still visible
  42
  43 die()
  44 {
  45         echo >&2
  46         echo "$*" >&2
  47         exit 1
  48 }
  49
  50 # When piped a commit, output a script to set the ident of either
  51 # "author" or "committer
  52
  53 set_ident () {
  54         lid="$(echo "$1" | tr "A-Z" "a-z")"
  55         uid="$(echo "$1" | tr "a-z" "A-Z")"
  56         pick_id_script='
  57                 /^'$lid' /{
  58                         s/'\''/'\''\\'\'\''/g
  59                         h
  60                         s/^'$lid' \([^<]*\) <[^>]*> .*$/\1/
  61                         s/'\''/'\''\'\'\''/g
  62                         s/.*/export GIT_'$uid'_NAME='\''&'\''/p
  63
  64                         g
  65                         s/^'$lid' [^<]* <\([^>]*\)> .*$/\1/
  66                         s/'\''/'\''\'\'\''/g
  67                         s/.*/export GIT_'$uid'_EMAIL='\''&'\''/p
  68
  69                         g
  70                         s/^'$lid' [^<]* <[^>]*> \(.*\)$/\1/
  71                         s/'\''/'\''\'\'\''/g
  72                         s/.*/export GIT_'$uid'_DATE='\''&'\''/p
  73
  74                         q
  75                 }
  76         '
  77
  78         LANG=C LC_ALL=C sed -ne "$pick_id_script"
  79         # Ensure non-empty id name.
  80         echo "[ -n \"\$GIT_${uid}_NAME\" ] || export GIT_${uid}_NAME=\"\${GIT_${uid}_EMAIL%%@*}\""
  81 }
  82
  83 # This script can be sourced by the commit filter to get the functions
  84 test "a$SOURCE_FUNCTIONS" = a1 && return
  85 this_script="$(cd "$(dirname "$0")"; pwd)"/$(basename "$0")
  86 export this_script
  87
  88 USAGE="[--env-filter <command>] [--tree-filter <command>] \
  89 [--index-filter <command>] [--parent-filter <command>] \
  90 [--msg-filter <command>] [--commit-filter <command>] \
  91 [--tag-name-filter <command>] [--subdirectory-filter <directory>] \
  92 [--original <namespace>] [-d <directory>] [-f | --force] \
  93 [<rev-list options>...]"
  94
  95 . git-sh-setup
  96
  97 git diff-files --quiet &&
  98         git diff-index --cached --quiet HEAD -- ||
  99         die "Cannot rewrite branch(es) with a dirty working directory."
 100
 101 tempdir=.git-rewrite
 102 filter_env=
 103 filter_tree=
 104 filter_index=
 105 filter_parent=
 106 filter_msg=cat
 107 filter_commit='git commit-tree "$@"'
 108 filter_tag_name=
 109 filter_subdir=
 110 orig_namespace=refs/original/
 111 force=
 112 while :
 113 do
 114         test $# = 0 && usage
 115         case "$1" in
 116         --)
 117                 shift
 118                 break
 119                 ;;
 120         --force|-f)
 121                 shift
 122                 force=t
 123                 continue
 124                 ;;
 125         -*)
 126                 ;;
 127         *)
 128                 break;
 129         esac
 130
 131         # all switches take one argument
 132         ARG="$1"
 133         case "$#" in 1) usage ;; esac
 134         shift
 135         OPTARG="$1"
 136         shift
 137
 138         case "$ARG" in
 139         -d)
 140                 tempdir="$OPTARG"
 141                 ;;
 142         --env-filter)
 143                 filter_env="$OPTARG"
 144                 ;;
 145         --tree-filter)
 146                 filter_tree="$OPTARG"
 147                 ;;
 148         --index-filter)
 149                 filter_index="$OPTARG"
 150                 ;;
 151         --parent-filter)
 152                 filter_parent="$OPTARG"
 153                 ;;
 154         --msg-filter)
 155                 filter_msg="$OPTARG"
 156                 ;;
 157         --commit-filter)
 158                 filter_commit='SOURCE_FUNCTIONS=1 . "$this_script";'" $OPTARG"
 159                 ;;
 160         --tag-name-filter)
 161                 filter_tag_name="$OPTARG"
 162                 ;;
 163         --subdirectory-filter)
 164                 filter_subdir="$OPTARG"
 165                 ;;
 166         --original)
 167                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 168                 ;;
 169         *)
 170                 usage
 171                 ;;
 172         esac
 173 done
 174
 175 case "$force" in
 176 t)
 177         rm -rf "$tempdir"
 178 ;;
 179 '')
 180         test -d "$tempdir" &&
 181                 die "$tempdir already exists, please remove it"
 182 esac
 183 mkdir -p "$tempdir/t" &&
 184 tempdir="$(cd "$tempdir"; pwd)" &&
 185 cd "$tempdir/t" &&
 186 workdir="$(pwd)" ||
 187 die ""
 188
 189 # Make sure refs/original is empty
 190 git for-each-ref > "$tempdir"/backup-refs
 191 while read sha1 type name
 192 do
 193         case "$force,$name" in
 194         ,$orig_namespace*)
 195                 die "Namespace $orig_namespace not empty"
 196         ;;
 197         t,$orig_namespace*)
 198                 git update-ref -d "$name" $sha1
 199         ;;
 200         esac
 201 done < "$tempdir"/backup-refs
 202
 203 ORIG_GIT_DIR="$GIT_DIR"
 204 ORIG_GIT_WORK_TREE="$GIT_WORK_TREE"
 205 ORIG_GIT_INDEX_FILE="$GIT_INDEX_FILE"
 206 export GIT_DIR GIT_WORK_TREE=.
 207
 208 # These refs should be updated if their heads were rewritten
 209
 210 git rev-parse --revs-only --symbolic "$@" |
 211 while read ref
 212 do
 213         # normalize ref
 214         case "$ref" in
 215         HEAD)
 216                 ref="$(git symbolic-ref "$ref")"
 217         ;;
 218         refs/*)
 219         ;;
 220         *)
 221                 ref="$(git for-each-ref --format='%(refname)' |
 222                         grep /"$ref")"
 223         esac
 224
 225         git check-ref-format "$ref" && echo "$ref"
 226 done > "$tempdir"/heads
 227
 228 test -s "$tempdir"/heads ||
 229         die "Which ref do you want to rewrite?"
 230
 231 export GIT_INDEX_FILE="$(pwd)/../index"
 232 git read-tree || die "Could not seed the index"
 233
 234 ret=0
 235
 236 # map old->new commit ids for rewriting parents
 237 mkdir ../map || die "Could not create map/ directory"
 238
 239 case "$filter_subdir" in
 240 "")
 241         git rev-list --reverse --topo-order --default HEAD \
 242                 --parents "$@"
 243         ;;
 244 *)
 245         git rev-list --reverse --topo-order --default HEAD \
 246                 --parents --full-history "$@" -- "$filter_subdir"
 247 esac > ../revs || die "Could not get the commits"
 248 commits=$(wc -l <../revs | tr -d " ")
 249
 250 test $commits -eq 0 && die "Found nothing to rewrite"
 251
 252 # Rewrite the commits
 253
 254 i=0
 255 while read commit parents; do
 256         i=$(($i+1))
 257         printf "\rRewrite $commit ($i/$commits)"
 258
 259         case "$filter_subdir" in
 260         "")
 261                 git read-tree -i -m $commit
 262                 ;;
 263         *)
 264                 git read-tree -i -m $commit:"$filter_subdir"
 265         esac || die "Could not initialize the index"
 266
 267         export GIT_COMMIT=$commit
 268         git cat-file commit "$commit" >../commit ||
 269                 die "Cannot read commit $commit"
 270
 271         eval "$(set_ident AUTHOR <../commit)" ||
 272                 die "setting author failed for commit $commit"
 273         eval "$(set_ident COMMITTER <../commit)" ||
 274                 die "setting committer failed for commit $commit"
 275         eval "$filter_env" < /dev/null ||
 276                 die "env filter failed: $filter_env"
 277
 278         if [ "$filter_tree" ]; then
 279                 git checkout-index -f -u -a ||
 280                         die "Could not checkout the index"
 281                 # files that $commit removed are now still in the working tree;
 282                 # remove them, else they would be added again
 283                 git ls-files -z --others | xargs -0 rm -f
 284                 eval "$filter_tree" < /dev/null ||
 285                         die "tree filter failed: $filter_tree"
 286
 287                 git diff-index -r $commit | cut -f 2- | tr '\n' '\0' | \
 288                         xargs -0 git update-index --add --replace --remove
 289                 git ls-files -z --others | \
 290                         xargs -0 git update-index --add --replace --remove
 291         fi
 292
 293         eval "$filter_index" < /dev/null ||
 294                 die "index filter failed: $filter_index"
 295
 296         parentstr=
 297         for parent in $parents; do
 298                 for reparent in $(map "$parent"); do
 299                         parentstr="$parentstr -p $reparent"
 300                 done
 301         done
 302         if [ "$filter_parent" ]; then
 303                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 304                                 die "parent filter failed: $filter_parent"
 305         fi
 306
 307         sed -e '1,/^$/d' <../commit | \
 308                 eval "$filter_msg" > ../message ||
 309                         die "msg filter failed: $filter_msg"
 310         sh -c "$filter_commit" "git commit-tree" \
 311                 $(git write-tree) $parentstr < ../message > ../map/$commit
 312 done <../revs
 313
 314 # In case of a subdirectory filter, it is possible that a specified head
 315 # is not in the set of rewritten commits, because it was pruned by the
 316 # revision walker.  Fix it by mapping these heads to the next rewritten
 317 # ancestor(s), i.e. the boundaries in the set of rewritten commits.
 318
 319 # NEEDSWORK: we should sort the unmapped refs topologically first
 320 while read ref
 321 do
 322         sha1=$(git rev-parse "$ref"^0)
 323         test -f "$workdir"/../map/$sha1 && continue
 324         # Assign the boundarie(s) in the set of rewritten commits
 325         # as the replacement commit(s).
 326         # (This would look a bit nicer if --not --stdin worked.)
 327         for p in $( (cd "$workdir"/../map; ls | sed "s/^/^/") |
 328                 git rev-list $ref --boundary --stdin |
 329                 sed -n "s/^-//p")
 330         do
 331                 map $p >> "$workdir"/../map/$sha1
 332         done
 333 done < "$tempdir"/heads
 334
 335 # Finally update the refs
 336
 337 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 338 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 339 count=0
 340 echo
 341 while read ref
 342 do
 343         # avoid rewriting a ref twice
 344         test -f "$orig_namespace$ref" && continue
 345
 346         sha1=$(git rev-parse "$ref"^0)
 347         rewritten=$(map $sha1)
 348
 349         test $sha1 = "$rewritten" &&
 350                 warn "WARNING: Ref '$ref' is unchanged" &&
 351                 continue
 352
 353         case "$rewritten" in
 354         '')
 355                 echo "Ref '$ref' was deleted"
 356                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 357                         die "Could not delete $ref"
 358         ;;
 359         $_x40)
 360                 echo "Ref '$ref' was rewritten"
 361                 git update-ref -m "filter-branch: rewrite" \
 362                                 "$ref" $rewritten $sha1 ||
 363                         die "Could not rewrite $ref"
 364         ;;
 365         *)
 366                 # NEEDSWORK: possibly add -Werror, making this an error
 367                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 368                 warn "$rewritten"
 369                 warn "WARNING: Ref '$ref' points to the first one now."
 370                 rewritten=$(echo "$rewritten" | head -n 1)
 371                 git update-ref -m "filter-branch: rewrite to first" \
 372                                 "$ref" $rewritten $sha1 ||
 373                         die "Could not rewrite $ref"
 374         ;;
 375         esac
 376         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1
 377         count=$(($count+1))
 378 done < "$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/$sha1")"
 401                 export GIT_COMMIT="$sha1"
 402                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 403                         die "tag name filter failed: $filter_tag_name"
 404
 405                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 406
 407                 if [ "$type" = "tag" ]; then
 408                         # Warn that we are not rewriting the tag object itself.
 409                         warn "unreferencing tag object $sha1t"
 410                 fi
 411
 412                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 413                         die "Could not write tag $new_ref"
 414         done
 415 fi
 416
 417 cd ../..
 418 rm -rf "$tempdir"
 419 echo
 420 test $count -gt 0 && echo "These refs were rewritten:"
 421 git show-ref | grep ^"$orig_namespace"
 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