contrib/subtree/git-subtree.sh

   1 #!/bin/bash
   2 #
   3 # git-subtree.sh: split/join git repositories in subdirectories of this one
   4 #
   5 # Copyright (C) 2009 Avery Pennarun <apenwarr@gmail.com>
   6 #
   7 if [ $# -eq 0 ]; then
   8     set -- -h
   9 fi
  10 OPTS_SPEC="\
  11 git subtree add   --prefix=<prefix> <commit>
  12 git subtree add   --prefix=<prefix> <repository> <commit>
  13 git subtree merge --prefix=<prefix> <commit>
  14 git subtree pull  --prefix=<prefix> <repository> <refspec...>
  15 git subtree push  --prefix=<prefix> <repository> <refspec...>
  16 git subtree split --prefix=<prefix> <commit...>
  17 --
  18 h,help        show the help
  19 q             quiet
  20 d             show debug messages
  21 P,prefix=     the name of the subdir to split out
  22 m,message=    use the given message as the commit message for the merge commit
  23  options for 'split'
  24 annotate=     add a prefix to commit message of new commits
  25 b,branch=     create a new branch from the split subtree
  26 ignore-joins  ignore prior --rejoin commits
  27 onto=         try connecting new tree to an existing one
  28 rejoin        merge the new branch back into HEAD
  29  options for 'add', 'merge', 'pull' and 'push'
  30 squash        merge subtree changes as a single commit
  31 "
  32 eval "$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@" || echo exit $?)"
  33
  34 PATH=$PATH:$(git --exec-path)
  35 . git-sh-setup
  36
  37 require_work_tree
  38
  39 quiet=
  40 branch=
  41 debug=
  42 command=
  43 onto=
  44 rejoin=
  45 ignore_joins=
  46 annotate=
  47 squash=
  48 message=
  49
  50 debug()
  51 {
  52         if [ -n "$debug" ]; then
  53                 echo "$@" >&2
  54         fi
  55 }
  56
  57 say()
  58 {
  59         if [ -z "$quiet" ]; then
  60                 echo "$@" >&2
  61         fi
  62 }
  63
  64 assert()
  65 {
  66         if "$@"; then
  67                 :
  68         else
  69                 die "assertion failed: " "$@"
  70         fi
  71 }
  72
  73
  74 #echo "Options: $*"
  75
  76 while [ $# -gt 0 ]; do
  77         opt="$1"
  78         shift
  79         case "$opt" in
  80                 -q) quiet=1 ;;
  81                 -d) debug=1 ;;
  82                 --annotate) annotate="$1"; shift ;;
  83                 --no-annotate) annotate= ;;
  84                 -b) branch="$1"; shift ;;
  85                 -P) prefix="$1"; shift ;;
  86                 -m) message="$1"; shift ;;
  87                 --no-prefix) prefix= ;;
  88                 --onto) onto="$1"; shift ;;
  89                 --no-onto) onto= ;;
  90                 --rejoin) rejoin=1 ;;
  91                 --no-rejoin) rejoin= ;;
  92                 --ignore-joins) ignore_joins=1 ;;
  93                 --no-ignore-joins) ignore_joins= ;;
  94                 --squash) squash=1 ;;
  95                 --no-squash) squash= ;;
  96                 --) break ;;
  97                 *) die "Unexpected option: $opt" ;;
  98         esac
  99 done
 100
 101 command="$1"
 102 shift
 103 case "$command" in
 104         add|merge|pull) default= ;;
 105         split|push) default="--default HEAD" ;;
 106         *) die "Unknown command '$command'" ;;
 107 esac
 108
 109 if [ -z "$prefix" ]; then
 110         die "You must provide the --prefix option."
 111 fi
 112
 113 case "$command" in
 114         add) [ -e "$prefix" ] &&
 115                 die "prefix '$prefix' already exists." ;;
 116         *)   [ -e "$prefix" ] ||
 117                 die "'$prefix' does not exist; use 'git subtree add'" ;;
 118 esac
 119
 120 dir="$(dirname "$prefix/.")"
 121
 122 if [ "$command" != "pull" -a "$command" != "add" -a "$command" != "push" ]; then
 123         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 124         dirs="$(git rev-parse --no-revs --no-flags "$@")" || exit $?
 125         if [ -n "$dirs" ]; then
 126                 die "Error: Use --prefix instead of bare filenames."
 127         fi
 128 fi
 129
 130 debug "command: {$command}"
 131 debug "quiet: {$quiet}"
 132 debug "revs: {$revs}"
 133 debug "dir: {$dir}"
 134 debug "opts: {$*}"
 135 debug
 136
 137 cache_setup()
 138 {
 139         cachedir="$GIT_DIR/subtree-cache/$$"
 140         rm -rf "$cachedir" || die "Can't delete old cachedir: $cachedir"
 141         mkdir -p "$cachedir" || die "Can't create new cachedir: $cachedir"
 142         mkdir -p "$cachedir/notree" || die "Can't create new cachedir: $cachedir/notree"
 143         debug "Using cachedir: $cachedir" >&2
 144 }
 145
 146 cache_get()
 147 {
 148         for oldrev in $*; do
 149                 if [ -r "$cachedir/$oldrev" ]; then
 150                         read newrev <"$cachedir/$oldrev"
 151                         echo $newrev
 152                 fi
 153         done
 154 }
 155
 156 cache_miss()
 157 {
 158         for oldrev in $*; do
 159                 if [ ! -r "$cachedir/$oldrev" ]; then
 160                         echo $oldrev
 161                 fi
 162         done
 163 }
 164
 165 check_parents()
 166 {
 167         missed=$(cache_miss $*)
 168         for miss in $missed; do
 169                 if [ ! -r "$cachedir/notree/$miss" ]; then
 170                         debug "  incorrect order: $miss"
 171                 fi
 172         done
 173 }
 174
 175 set_notree()
 176 {
 177         echo "1" > "$cachedir/notree/$1"
 178 }
 179
 180 cache_set()
 181 {
 182         oldrev="$1"
 183         newrev="$2"
 184         if [ "$oldrev" != "latest_old" \
 185              -a "$oldrev" != "latest_new" \
 186              -a -e "$cachedir/$oldrev" ]; then
 187                 die "cache for $oldrev already exists!"
 188         fi
 189         echo "$newrev" >"$cachedir/$oldrev"
 190 }
 191
 192 rev_exists()
 193 {
 194         if git rev-parse "$1" >/dev/null 2>&1; then
 195                 return 0
 196         else
 197                 return 1
 198         fi
 199 }
 200
 201 rev_is_descendant_of_branch()
 202 {
 203         newrev="$1"
 204         branch="$2"
 205         branch_hash=$(git rev-parse $branch)
 206         match=$(git rev-list -1 $branch_hash ^$newrev)
 207
 208         if [ -z "$match" ]; then
 209                 return 0
 210         else
 211                 return 1
 212         fi
 213 }
 214
 215 # if a commit doesn't have a parent, this might not work.  But we only want
 216 # to remove the parent from the rev-list, and since it doesn't exist, it won't
 217 # be there anyway, so do nothing in that case.
 218 try_remove_previous()
 219 {
 220         if rev_exists "$1^"; then
 221                 echo "^$1^"
 222         fi
 223 }
 224
 225 find_latest_squash()
 226 {
 227         debug "Looking for latest squash ($dir)..."
 228         dir="$1"
 229         sq=
 230         main=
 231         sub=
 232         git log --grep="^git-subtree-dir: $dir/*\$" \
 233                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' HEAD |
 234         while read a b junk; do
 235                 debug "$a $b $junk"
 236                 debug "{{$sq/$main/$sub}}"
 237                 case "$a" in
 238                         START) sq="$b" ;;
 239                         git-subtree-mainline:) main="$b" ;;
 240                         git-subtree-split:) sub="$b" ;;
 241                         END)
 242                                 if [ -n "$sub" ]; then
 243                                         if [ -n "$main" ]; then
 244                                                 # a rejoin commit?
 245                                                 # Pretend its sub was a squash.
 246                                                 sq="$sub"
 247                                         fi
 248                                         debug "Squash found: $sq $sub"
 249                                         echo "$sq" "$sub"
 250                                         break
 251                                 fi
 252                                 sq=
 253                                 main=
 254                                 sub=
 255                                 ;;
 256                 esac
 257         done
 258 }
 259
 260 find_existing_splits()
 261 {
 262         debug "Looking for prior splits..."
 263         dir="$1"
 264         revs="$2"
 265         main=
 266         sub=
 267         git log --grep="^git-subtree-dir: $dir/*\$" \
 268                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' $revs |
 269         while read a b junk; do
 270                 case "$a" in
 271                         START) sq="$b" ;;
 272                         git-subtree-mainline:) main="$b" ;;
 273                         git-subtree-split:) sub="$b" ;;
 274                         END)
 275                                 debug "  Main is: '$main'"
 276                                 if [ -z "$main" -a -n "$sub" ]; then
 277                                         # squash commits refer to a subtree
 278                                         debug "  Squash: $sq from $sub"
 279                                         cache_set "$sq" "$sub"
 280                                 fi
 281                                 if [ -n "$main" -a -n "$sub" ]; then
 282                                         debug "  Prior: $main -> $sub"
 283                                         cache_set $main $sub
 284                                         cache_set $sub $sub
 285                                         try_remove_previous "$main"
 286                                         try_remove_previous "$sub"
 287                                 fi
 288                                 main=
 289                                 sub=
 290                                 ;;
 291                 esac
 292         done
 293 }
 294
 295 copy_commit()
 296 {
 297         # We're going to set some environment vars here, so
 298         # do it in a subshell to get rid of them safely later
 299         debug copy_commit "{$1}" "{$2}" "{$3}"
 300         git log -1 --pretty=format:'%an%n%ae%n%ad%n%cn%n%ce%n%cd%n%B' "$1" |
 301         (
 302                 read GIT_AUTHOR_NAME
 303                 read GIT_AUTHOR_EMAIL
 304                 read GIT_AUTHOR_DATE
 305                 read GIT_COMMITTER_NAME
 306                 read GIT_COMMITTER_EMAIL
 307                 read GIT_COMMITTER_DATE
 308                 export  GIT_AUTHOR_NAME \
 309                         GIT_AUTHOR_EMAIL \
 310                         GIT_AUTHOR_DATE \
 311                         GIT_COMMITTER_NAME \
 312                         GIT_COMMITTER_EMAIL \
 313                         GIT_COMMITTER_DATE
 314                 (printf "%s" "$annotate"; cat ) |
 315                 git commit-tree "$2" $3  # reads the rest of stdin
 316         ) || die "Can't copy commit $1"
 317 }
 318
 319 add_msg()
 320 {
 321         dir="$1"
 322         latest_old="$2"
 323         latest_new="$3"
 324         if [ -n "$message" ]; then
 325                 commit_message="$message"
 326         else
 327                 commit_message="Add '$dir/' from commit '$latest_new'"
 328         fi
 329         cat <<-EOF
 330                 $commit_message
 331
 332                 git-subtree-dir: $dir
 333                 git-subtree-mainline: $latest_old
 334                 git-subtree-split: $latest_new
 335         EOF
 336 }
 337
 338 add_squashed_msg()
 339 {
 340         if [ -n "$message" ]; then
 341                 echo "$message"
 342         else
 343                 echo "Merge commit '$1' as '$2'"
 344         fi
 345 }
 346
 347 rejoin_msg()
 348 {
 349         dir="$1"
 350         latest_old="$2"
 351         latest_new="$3"
 352         if [ -n "$message" ]; then
 353                 commit_message="$message"
 354         else
 355                 commit_message="Split '$dir/' into commit '$latest_new'"
 356         fi
 357         cat <<-EOF
 358                 $commit_message
 359
 360                 git-subtree-dir: $dir
 361                 git-subtree-mainline: $latest_old
 362                 git-subtree-split: $latest_new
 363         EOF
 364 }
 365
 366 squash_msg()
 367 {
 368         dir="$1"
 369         oldsub="$2"
 370         newsub="$3"
 371         newsub_short=$(git rev-parse --short "$newsub")
 372
 373         if [ -n "$oldsub" ]; then
 374                 oldsub_short=$(git rev-parse --short "$oldsub")
 375                 echo "Squashed '$dir/' changes from $oldsub_short..$newsub_short"
 376                 echo
 377                 git log --pretty=tformat:'%h %s' "$oldsub..$newsub"
 378                 git log --pretty=tformat:'REVERT: %h %s' "$newsub..$oldsub"
 379         else
 380                 echo "Squashed '$dir/' content from commit $newsub_short"
 381         fi
 382
 383         echo
 384         echo "git-subtree-dir: $dir"
 385         echo "git-subtree-split: $newsub"
 386 }
 387
 388 toptree_for_commit()
 389 {
 390         commit="$1"
 391         git log -1 --pretty=format:'%T' "$commit" -- || exit $?
 392 }
 393
 394 subtree_for_commit()
 395 {
 396         commit="$1"
 397         dir="$2"
 398         git ls-tree "$commit" -- "$dir" |
 399         while read mode type tree name; do
 400                 assert [ "$name" = "$dir" ]
 401                 assert [ "$type" = "tree" -o "$type" = "commit" ]
 402                 [ "$type" = "commit" ] && continue  # ignore submodules
 403                 echo $tree
 404                 break
 405         done
 406 }
 407
 408 tree_changed()
 409 {
 410         tree=$1
 411         shift
 412         if [ $# -ne 1 ]; then
 413                 return 0   # weird parents, consider it changed
 414         else
 415                 ptree=$(toptree_for_commit $1)
 416                 if [ "$ptree" != "$tree" ]; then
 417                         return 0   # changed
 418                 else
 419                         return 1   # not changed
 420                 fi
 421         fi
 422 }
 423
 424 new_squash_commit()
 425 {
 426         old="$1"
 427         oldsub="$2"
 428         newsub="$3"
 429         tree=$(toptree_for_commit $newsub) || exit $?
 430         if [ -n "$old" ]; then
 431                 squash_msg "$dir" "$oldsub" "$newsub" |
 432                         git commit-tree "$tree" -p "$old" || exit $?
 433         else
 434                 squash_msg "$dir" "" "$newsub" |
 435                         git commit-tree "$tree" || exit $?
 436         fi
 437 }
 438
 439 copy_or_skip()
 440 {
 441         rev="$1"
 442         tree="$2"
 443         newparents="$3"
 444         assert [ -n "$tree" ]
 445
 446         identical=
 447         nonidentical=
 448         p=
 449         gotparents=
 450         for parent in $newparents; do
 451                 ptree=$(toptree_for_commit $parent) || exit $?
 452                 [ -z "$ptree" ] && continue
 453                 if [ "$ptree" = "$tree" ]; then
 454                         # an identical parent could be used in place of this rev.
 455                         identical="$parent"
 456                 else
 457                         nonidentical="$parent"
 458                 fi
 459
 460                 # sometimes both old parents map to the same newparent;
 461                 # eliminate duplicates
 462                 is_new=1
 463                 for gp in $gotparents; do
 464                         if [ "$gp" = "$parent" ]; then
 465                                 is_new=
 466                                 break
 467                         fi
 468                 done
 469                 if [ -n "$is_new" ]; then
 470                         gotparents="$gotparents $parent"
 471                         p="$p -p $parent"
 472                 fi
 473         done
 474
 475         if [ -n "$identical" ]; then
 476                 echo $identical
 477         else
 478                 copy_commit $rev $tree "$p" || exit $?
 479         fi
 480 }
 481
 482 ensure_clean()
 483 {
 484         if ! git diff-index HEAD --exit-code --quiet 2>&1; then
 485                 die "Working tree has modifications.  Cannot add."
 486         fi
 487         if ! git diff-index --cached HEAD --exit-code --quiet 2>&1; then
 488                 die "Index has modifications.  Cannot add."
 489         fi
 490 }
 491
 492 cmd_add()
 493 {
 494         if [ -e "$dir" ]; then
 495                 die "'$dir' already exists.  Cannot add."
 496         fi
 497
 498         ensure_clean
 499
 500         if [ $# -eq 1 ]; then
 501             git rev-parse -q --verify "$1^{commit}" >/dev/null ||
 502             die "'$1' does not refer to a commit"
 503
 504             "cmd_add_commit" "$@"
 505         elif [ $# -eq 2 ]; then
 506             # Technically we could accept a refspec here but we're
 507             # just going to turn around and add FETCH_HEAD under the
 508             # specified directory.  Allowing a refspec might be
 509             # misleading because we won't do anything with any other
 510             # branches fetched via the refspec.
 511             git rev-parse -q --verify "$2^{commit}" >/dev/null ||
 512             die "'$2' does not refer to a commit"
 513
 514             "cmd_add_repository" "$@"
 515         else
 516             say "error: parameters were '$@'"
 517             die "Provide either a commit or a repository and commit."
 518         fi
 519 }
 520
 521 cmd_add_repository()
 522 {
 523         echo "git fetch" "$@"
 524         repository=$1
 525         refspec=$2
 526         git fetch "$@" || exit $?
 527         revs=FETCH_HEAD
 528         set -- $revs
 529         cmd_add_commit "$@"
 530 }
 531
 532 cmd_add_commit()
 533 {
 534         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 535         set -- $revs
 536         rev="$1"
 537
 538         debug "Adding $dir as '$rev'..."
 539         git read-tree --prefix="$dir" $rev || exit $?
 540         git checkout -- "$dir" || exit $?
 541         tree=$(git write-tree) || exit $?
 542
 543         headrev=$(git rev-parse HEAD) || exit $?
 544         if [ -n "$headrev" -a "$headrev" != "$rev" ]; then
 545                 headp="-p $headrev"
 546         else
 547                 headp=
 548         fi
 549
 550         if [ -n "$squash" ]; then
 551                 rev=$(new_squash_commit "" "" "$rev") || exit $?
 552                 commit=$(add_squashed_msg "$rev" "$dir" |
 553                          git commit-tree $tree $headp -p "$rev") || exit $?
 554         else
 555                 commit=$(add_msg "$dir" "$headrev" "$rev" |
 556                          git commit-tree $tree $headp -p "$rev") || exit $?
 557         fi
 558         git reset "$commit" || exit $?
 559
 560         say "Added dir '$dir'"
 561 }
 562
 563 cmd_split()
 564 {
 565         debug "Splitting $dir..."
 566         cache_setup || exit $?
 567
 568         if [ -n "$onto" ]; then
 569                 debug "Reading history for --onto=$onto..."
 570                 git rev-list $onto |
 571                 while read rev; do
 572                         # the 'onto' history is already just the subdir, so
 573                         # any parent we find there can be used verbatim
 574                         debug "  cache: $rev"
 575                         cache_set $rev $rev
 576                 done
 577         fi
 578
 579         if [ -n "$ignore_joins" ]; then
 580                 unrevs=
 581         else
 582                 unrevs="$(find_existing_splits "$dir" "$revs")"
 583         fi
 584
 585         # We can't restrict rev-list to only $dir here, because some of our
 586         # parents have the $dir contents the root, and those won't match.
 587         # (and rev-list --follow doesn't seem to solve this)
 588         grl='git rev-list --topo-order --reverse --parents $revs $unrevs'
 589         revmax=$(eval "$grl" | wc -l)
 590         revcount=0
 591         createcount=0
 592         eval "$grl" |
 593         while read rev parents; do
 594                 revcount=$(($revcount + 1))
 595                 say -n "$revcount/$revmax ($createcount)
"
 596                 debug "Processing commit: $rev"
 597                 exists=$(cache_get $rev)
 598                 if [ -n "$exists" ]; then
 599                         debug "  prior: $exists"
 600                         continue
 601                 fi
 602                 createcount=$(($createcount + 1))
 603                 debug "  parents: $parents"
 604                 newparents=$(cache_get $parents)
 605                 debug "  newparents: $newparents"
 606
 607                 tree=$(subtree_for_commit $rev "$dir")
 608                 debug "  tree is: $tree"
 609
 610                 check_parents $parents
 611
 612                 # ugly.  is there no better way to tell if this is a subtree
 613                 # vs. a mainline commit?  Does it matter?
 614                 if [ -z $tree ]; then
 615                         set_notree $rev
 616                         if [ -n "$newparents" ]; then
 617                                 cache_set $rev $rev
 618                         fi
 619                         continue
 620                 fi
 621
 622                 newrev=$(copy_or_skip "$rev" "$tree" "$newparents") || exit $?
 623                 debug "  newrev is: $newrev"
 624                 cache_set $rev $newrev
 625                 cache_set latest_new $newrev
 626                 cache_set latest_old $rev
 627         done || exit $?
 628         latest_new=$(cache_get latest_new)
 629         if [ -z "$latest_new" ]; then
 630                 die "No new revisions were found"
 631         fi
 632
 633         if [ -n "$rejoin" ]; then
 634                 debug "Merging split branch into HEAD..."
 635                 latest_old=$(cache_get latest_old)
 636                 git merge -s ours \
 637                         -m "$(rejoin_msg $dir $latest_old $latest_new)" \
 638                         $latest_new >&2 || exit $?
 639         fi
 640         if [ -n "$branch" ]; then
 641                 if rev_exists "refs/heads/$branch"; then
 642                         if ! rev_is_descendant_of_branch $latest_new $branch; then
 643                                 die "Branch '$branch' is not an ancestor of commit '$latest_new'."
 644                         fi
 645                         action='Updated'
 646                 else
 647                         action='Created'
 648                 fi
 649                 git update-ref -m 'subtree split' "refs/heads/$branch" $latest_new || exit $?
 650                 say "$action branch '$branch'"
 651         fi
 652         echo $latest_new
 653         exit 0
 654 }
 655
 656 cmd_merge()
 657 {
 658         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 659         ensure_clean
 660
 661         set -- $revs
 662         if [ $# -ne 1 ]; then
 663                 die "You must provide exactly one revision.  Got: '$revs'"
 664         fi
 665         rev="$1"
 666
 667         if [ -n "$squash" ]; then
 668                 first_split="$(find_latest_squash "$dir")"
 669                 if [ -z "$first_split" ]; then
 670                         die "Can't squash-merge: '$dir' was never added."
 671                 fi
 672                 set $first_split
 673                 old=$1
 674                 sub=$2
 675                 if [ "$sub" = "$rev" ]; then
 676                         say "Subtree is already at commit $rev."
 677                         exit 0
 678                 fi
 679                 new=$(new_squash_commit "$old" "$sub" "$rev") || exit $?
 680                 debug "New squash commit: $new"
 681                 rev="$new"
 682         fi
 683
 684         version=$(git version)
 685         if [ "$version" \< "git version 1.7" ]; then
 686                 if [ -n "$message" ]; then
 687                         git merge -s subtree --message="$message" $rev
 688                 else
 689                         git merge -s subtree $rev
 690                 fi
 691         else
 692                 if [ -n "$message" ]; then
 693                         git merge -Xsubtree="$prefix" --message="$message" $rev
 694                 else
 695                         git merge -Xsubtree="$prefix" $rev
 696                 fi
 697         fi
 698 }
 699
 700 cmd_pull()
 701 {
 702         ensure_clean
 703         git fetch "$@" || exit $?
 704         revs=FETCH_HEAD
 705         set -- $revs
 706         cmd_merge "$@"
 707 }
 708
 709 cmd_push()
 710 {
 711         if [ $# -ne 2 ]; then
 712             die "You must provide <repository> <refspec>"
 713         fi
 714         if [ -e "$dir" ]; then
 715             repository=$1
 716             refspec=$2
 717             echo "git push using: " $repository $refspec
 718             git push $repository $(git subtree split --prefix=$prefix):refs/heads/$refspec
 719         else
 720             die "'$dir' must already exist. Try 'git subtree add'."
 721         fi
 722 }
 723
 724 "cmd_$command" "$@"