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 "$1"
  23         fi
  24 }
  25
  26 # override die(): this version puts in an extra line break, so that
  27 # the progress is still visible
  28
  29 die()
  30 {
  31         echo >&2
  32         echo "$*" >&2
  33         exit 1
  34 }
  35
  36 # When piped a commit, output a script to set the ident of either
  37 # "author" or "committer
  38
  39 set_ident () {
  40         lid="$(echo "$1" | tr "A-Z" "a-z")"
  41         uid="$(echo "$1" | tr "a-z" "A-Z")"
  42         pick_id_script='
  43                 /^'$lid' /{
  44                         s/'\''/'\''\\'\'\''/g
  45                         h
  46                         s/^'$lid' \([^<]*\) <[^>]*> .*$/\1/
  47                         s/'\''/'\''\'\'\''/g
  48                         s/.*/export GIT_'$uid'_NAME='\''&'\''/p
  49
  50                         g
  51                         s/^'$lid' [^<]* <\([^>]*\)> .*$/\1/
  52                         s/'\''/'\''\'\'\''/g
  53                         s/.*/export GIT_'$uid'_EMAIL='\''&'\''/p
  54
  55                         g
  56                         s/^'$lid' [^<]* <[^>]*> \(.*\)$/\1/
  57                         s/'\''/'\''\'\'\''/g
  58                         s/.*/export GIT_'$uid'_DATE='\''&'\''/p
  59
  60                         q
  61                 }
  62         '
  63
  64         LANG=C LC_ALL=C sed -ne "$pick_id_script"
  65         # Ensure non-empty id name.
  66         echo "[ -n \"\$GIT_${uid}_NAME\" ] || export GIT_${uid}_NAME=\"\${GIT_${uid}_EMAIL%%@*}\""
  67 }
  68
  69 # This script can be sourced by the commit filter to get the functions
  70 test "a$SOURCE_FUNCTIONS" = a1 && return
  71 this_script="$(cd "$(dirname "$0")"; pwd)"/$(basename "$0")
  72 export this_script
  73
  74 USAGE="[--env-filter <command>] [--tree-filter <command>] \
  75 [--index-filter <command>] [--parent-filter <command>] \
  76 [--msg-filter <command>] [--commit-filter <command>] \
  77 [--tag-name-filter <command>] [--subdirectory-filter <directory>] \
  78 [--original <namespace>] [-d <directory>] [-f | --force] \
  79 [<rev-list options>...]"
  80
  81 . git-sh-setup
  82
  83 tempdir=.git-rewrite
  84 filter_env=
  85 filter_tree=
  86 filter_index=
  87 filter_parent=
  88 filter_msg=cat
  89 filter_commit='git commit-tree "$@"'
  90 filter_tag_name=
  91 filter_subdir=
  92 orig_namespace=refs/original/
  93 force=
  94 while case "$#" in 0) usage;; esac
  95 do
  96         case "$1" in
  97         --)
  98                 shift
  99                 break
 100                 ;;
 101         --force|-f)
 102                 shift
 103                 force=t
 104                 continue
 105                 ;;
 106         -*)
 107                 ;;
 108         *)
 109                 break;
 110         esac
 111
 112         # all switches take one argument
 113         ARG="$1"
 114         case "$#" in 1) usage ;; esac
 115         shift
 116         OPTARG="$1"
 117         shift
 118
 119         case "$ARG" in
 120         -d)
 121                 tempdir="$OPTARG"
 122                 ;;
 123         --env-filter)
 124                 filter_env="$OPTARG"
 125                 ;;
 126         --tree-filter)
 127                 filter_tree="$OPTARG"
 128                 ;;
 129         --index-filter)
 130                 filter_index="$OPTARG"
 131                 ;;
 132         --parent-filter)
 133                 filter_parent="$OPTARG"
 134                 ;;
 135         --msg-filter)
 136                 filter_msg="$OPTARG"
 137                 ;;
 138         --commit-filter)
 139                 filter_commit='SOURCE_FUNCTIONS=1 . "$this_script";'" $OPTARG"
 140                 ;;
 141         --tag-name-filter)
 142                 filter_tag_name="$OPTARG"
 143                 ;;
 144         --subdirectory-filter)
 145                 filter_subdir="$OPTARG"
 146                 ;;
 147         --original)
 148                 orig_namespace=$(expr "$OPTARG/" : '\(.*[^/]\)/*$')/
 149                 ;;
 150         *)
 151                 usage
 152                 ;;
 153         esac
 154 done
 155
 156 case "$force" in
 157 t)
 158         rm -rf "$tempdir"
 159 ;;
 160 '')
 161         test -d "$tempdir" &&
 162                 die "$tempdir already exists, please remove it"
 163 esac
 164 mkdir -p "$tempdir/t" &&
 165 tempdir="$(cd "$tempdir"; pwd)" &&
 166 cd "$tempdir/t" &&
 167 workdir="$(pwd)" ||
 168 die ""
 169
 170 # Make sure refs/original is empty
 171 git for-each-ref > "$tempdir"/backup-refs
 172 while read sha1 type name
 173 do
 174         case "$force,$name" in
 175         ,$orig_namespace*)
 176                 die "Namespace $orig_namespace not empty"
 177         ;;
 178         t,$orig_namespace*)
 179                 git update-ref -d "$name" $sha1
 180         ;;
 181         esac
 182 done < "$tempdir"/backup-refs
 183
 184 export GIT_DIR GIT_WORK_TREE=.
 185
 186 # These refs should be updated if their heads were rewritten
 187
 188 git rev-parse --revs-only --symbolic "$@" |
 189 while read ref
 190 do
 191         # normalize ref
 192         case "$ref" in
 193         HEAD)
 194                 ref="$(git symbolic-ref "$ref")"
 195         ;;
 196         refs/*)
 197         ;;
 198         *)
 199                 ref="$(git for-each-ref --format='%(refname)' |
 200                         grep /"$ref")"
 201         esac
 202
 203         git check-ref-format "$ref" && echo "$ref"
 204 done > "$tempdir"/heads
 205
 206 test -s "$tempdir"/heads ||
 207         die "Which ref do you want to rewrite?"
 208
 209 export GIT_INDEX_FILE="$(pwd)/../index"
 210 git read-tree || die "Could not seed the index"
 211
 212 ret=0
 213
 214 # map old->new commit ids for rewriting parents
 215 mkdir ../map || die "Could not create map/ directory"
 216
 217 case "$filter_subdir" in
 218 "")
 219         git rev-list --reverse --topo-order --default HEAD \
 220                 --parents "$@"
 221         ;;
 222 *)
 223         git rev-list --reverse --topo-order --default HEAD \
 224                 --parents --full-history "$@" -- "$filter_subdir"
 225 esac > ../revs || die "Could not get the commits"
 226 commits=$(wc -l <../revs | tr -d " ")
 227
 228 test $commits -eq 0 && die "Found nothing to rewrite"
 229
 230 # Rewrite the commits
 231
 232 i=0
 233 while read commit parents; do
 234         i=$(($i+1))
 235         printf "\rRewrite $commit ($i/$commits)"
 236
 237         case "$filter_subdir" in
 238         "")
 239                 git read-tree -i -m $commit
 240                 ;;
 241         *)
 242                 git read-tree -i -m $commit:"$filter_subdir"
 243         esac || die "Could not initialize the index"
 244
 245         export GIT_COMMIT=$commit
 246         git cat-file commit "$commit" >../commit ||
 247                 die "Cannot read commit $commit"
 248
 249         eval "$(set_ident AUTHOR <../commit)" ||
 250                 die "setting author failed for commit $commit"
 251         eval "$(set_ident COMMITTER <../commit)" ||
 252                 die "setting committer failed for commit $commit"
 253         eval "$filter_env" < /dev/null ||
 254                 die "env filter failed: $filter_env"
 255
 256         if [ "$filter_tree" ]; then
 257                 git checkout-index -f -u -a ||
 258                         die "Could not checkout the index"
 259                 # files that $commit removed are now still in the working tree;
 260                 # remove them, else they would be added again
 261                 git ls-files -z --others | xargs -0 rm -f
 262                 eval "$filter_tree" < /dev/null ||
 263                         die "tree filter failed: $filter_tree"
 264
 265                 git diff-index -r $commit | cut -f 2- | tr '\n' '\0' | \
 266                         xargs -0 git update-index --add --replace --remove
 267                 git ls-files -z --others | \
 268                         xargs -0 git update-index --add --replace --remove
 269         fi
 270
 271         eval "$filter_index" < /dev/null ||
 272                 die "index filter failed: $filter_index"
 273
 274         parentstr=
 275         for parent in $parents; do
 276                 for reparent in $(map "$parent"); do
 277                         parentstr="$parentstr -p $reparent"
 278                 done
 279         done
 280         if [ "$filter_parent" ]; then
 281                 parentstr="$(echo "$parentstr" | eval "$filter_parent")" ||
 282                                 die "parent filter failed: $filter_parent"
 283         fi
 284
 285         sed -e '1,/^$/d' <../commit | \
 286                 eval "$filter_msg" > ../message ||
 287                         die "msg filter failed: $filter_msg"
 288         sh -c "$filter_commit" "git commit-tree" \
 289                 $(git write-tree) $parentstr < ../message > ../map/$commit
 290 done <../revs
 291
 292 # In case of a subdirectory filter, it is possible that a specified head
 293 # is not in the set of rewritten commits, because it was pruned by the
 294 # revision walker.  Fix it by mapping these heads to the next rewritten
 295 # ancestor(s), i.e. the boundaries in the set of rewritten commits.
 296
 297 # NEEDSWORK: we should sort the unmapped refs topologically first
 298 while read ref
 299 do
 300         sha1=$(git rev-parse "$ref"^0)
 301         test -f "$workdir"/../map/$sha1 && continue
 302         # Assign the boundarie(s) in the set of rewritten commits
 303         # as the replacement commit(s).
 304         # (This would look a bit nicer if --not --stdin worked.)
 305         for p in $( (cd "$workdir"/../map; ls | sed "s/^/^/") |
 306                 git rev-list $ref --boundary --stdin |
 307                 sed -n "s/^-//p")
 308         do
 309                 map $p >> "$workdir"/../map/$sha1
 310         done
 311 done < "$tempdir"/heads
 312
 313 # Finally update the refs
 314
 315 _x40='[0-9a-f][0-9a-f][0-9a-f][0-9a-f][0-9a-f]'
 316 _x40="$_x40$_x40$_x40$_x40$_x40$_x40$_x40$_x40"
 317 count=0
 318 echo
 319 while read ref
 320 do
 321         # avoid rewriting a ref twice
 322         test -f "$orig_namespace$ref" && continue
 323
 324         sha1=$(git rev-parse "$ref"^0)
 325         rewritten=$(map $sha1)
 326
 327         test $sha1 = "$rewritten" &&
 328                 warn "WARNING: Ref '$ref' is unchanged" &&
 329                 continue
 330
 331         case "$rewritten" in
 332         '')
 333                 echo "Ref '$ref' was deleted"
 334                 git update-ref -m "filter-branch: delete" -d "$ref" $sha1 ||
 335                         die "Could not delete $ref"
 336         ;;
 337         $_x40)
 338                 echo "Ref '$ref' was rewritten"
 339                 git update-ref -m "filter-branch: rewrite" \
 340                                 "$ref" $rewritten $sha1 ||
 341                         die "Could not rewrite $ref"
 342         ;;
 343         *)
 344                 # NEEDSWORK: possibly add -Werror, making this an error
 345                 warn "WARNING: '$ref' was rewritten into multiple commits:"
 346                 warn "$rewritten"
 347                 warn "WARNING: Ref '$ref' points to the first one now."
 348                 rewritten=$(echo "$rewritten" | head -n 1)
 349                 git update-ref -m "filter-branch: rewrite to first" \
 350                                 "$ref" $rewritten $sha1 ||
 351                         die "Could not rewrite $ref"
 352         ;;
 353         esac
 354         git update-ref -m "filter-branch: backup" "$orig_namespace$ref" $sha1
 355         count=$(($count+1))
 356 done < "$tempdir"/heads
 357
 358 # TODO: This should possibly go, with the semantics that all positive given
 359 #       refs are updated, and their original heads stored in refs/original/
 360 # Filter tags
 361
 362 if [ "$filter_tag_name" ]; then
 363         git for-each-ref --format='%(objectname) %(objecttype) %(refname)' refs/tags |
 364         while read sha1 type ref; do
 365                 ref="${ref#refs/tags/}"
 366                 # XXX: Rewrite tagged trees as well?
 367                 if [ "$type" != "commit" -a "$type" != "tag" ]; then
 368                         continue;
 369                 fi
 370
 371                 if [ "$type" = "tag" ]; then
 372                         # Dereference to a commit
 373                         sha1t="$sha1"
 374                         sha1="$(git rev-parse "$sha1"^{commit} 2>/dev/null)" || continue
 375                 fi
 376
 377                 [ -f "../map/$sha1" ] || continue
 378                 new_sha1="$(cat "../map/$sha1")"
 379                 export GIT_COMMIT="$sha1"
 380                 new_ref="$(echo "$ref" | eval "$filter_tag_name")" ||
 381                         die "tag name filter failed: $filter_tag_name"
 382
 383                 echo "$ref -> $new_ref ($sha1 -> $new_sha1)"
 384
 385                 if [ "$type" = "tag" ]; then
 386                         # Warn that we are not rewriting the tag object itself.
 387                         warn "unreferencing tag object $sha1t"
 388                 fi
 389
 390                 git update-ref "refs/tags/$new_ref" "$new_sha1" ||
 391                         die "Could not write tag $new_ref"
 392         done
 393 fi
 394
 395 cd ../..
 396 rm -rf "$tempdir"
 397 echo
 398 test $count -gt 0 && echo "These refs were rewritten:"
 399 git show-ref | grep ^"$orig_namespace"
 400
 401 exit $ret