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