contrib/subtree/git-subtree.sh

   1 #!/bin/sh
   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 [ $# -eq 0 ]; then
   8     set -- -h
   9 fi
  10 OPTS_SPEC="\
  11 git subtree add   --prefix=<prefix> <commit>
  12 git subtree add   --prefix=<prefix> <repository> <ref>
  13 git subtree merge --prefix=<prefix> <commit>
  14 git subtree pull  --prefix=<prefix> <repository> <ref>
  15 git subtree push  --prefix=<prefix> <repository> <ref>
  16 git subtree split --prefix=<prefix> <commit...>
  17 --
  18 h,help        show the help
  19 q             quiet
  20 d             show debug messages
  21 P,prefix=     the name of the subdir to split out
  22 m,message=    use the given message as the commit message for the merge commit
  23  options for 'split'
  24 annotate=     add a prefix to commit message of new commits
  25 b,branch=     create a new branch from the split subtree
  26 ignore-joins  ignore prior --rejoin commits
  27 onto=         try connecting new tree to an existing one
  28 rejoin        merge the new branch back into HEAD
  29  options for 'add', 'merge', 'pull' and 'push'
  30 squash        merge subtree changes as a single commit
  31 "
  32 eval "$(echo "$OPTS_SPEC" | git rev-parse --parseopt -- "$@" || echo exit $?)"
  33
  34 PATH=$PATH:$(git --exec-path)
  35 . git-sh-setup
  36
  37 require_work_tree
  38
  39 quiet=
  40 branch=
  41 debug=
  42 command=
  43 onto=
  44 rejoin=
  45 ignore_joins=
  46 annotate=
  47 squash=
  48 message=
  49 prefix=
  50
  51 debug()
  52 {
  53         if [ -n "$debug" ]; then
  54                 echo "$@" >&2
  55         fi
  56 }
  57
  58 say()
  59 {
  60         if [ -z "$quiet" ]; then
  61                 echo "$@" >&2
  62         fi
  63 }
  64
  65 assert()
  66 {
  67         if "$@"; then
  68                 :
  69         else
  70                 die "assertion failed: " "$@"
  71         fi
  72 }
  73
  74
  75 #echo "Options: $*"
  76
  77 while [ $# -gt 0 ]; do
  78         opt="$1"
  79         shift
  80         case "$opt" in
  81                 -q) quiet=1 ;;
  82                 -d) debug=1 ;;
  83                 --annotate) annotate="$1"; shift ;;
  84                 --no-annotate) annotate= ;;
  85                 -b) branch="$1"; shift ;;
  86                 -P) prefix="$1"; shift ;;
  87                 -m) message="$1"; shift ;;
  88                 --no-prefix) prefix= ;;
  89                 --onto) onto="$1"; shift ;;
  90                 --no-onto) onto= ;;
  91                 --rejoin) rejoin=1 ;;
  92                 --no-rejoin) rejoin= ;;
  93                 --ignore-joins) ignore_joins=1 ;;
  94                 --no-ignore-joins) ignore_joins= ;;
  95                 --squash) squash=1 ;;
  96                 --no-squash) squash= ;;
  97                 --) break ;;
  98                 *) die "Unexpected option: $opt" ;;
  99         esac
 100 done
 101
 102 command="$1"
 103 shift
 104 case "$command" in
 105         add|merge|pull) default= ;;
 106         split|push) default="--default HEAD" ;;
 107         *) die "Unknown command '$command'" ;;
 108 esac
 109
 110 if [ -z "$prefix" ]; then
 111         die "You must provide the --prefix option."
 112 fi
 113
 114 case "$command" in
 115         add) [ -e "$prefix" ] &&
 116                 die "prefix '$prefix' already exists." ;;
 117         *)   [ -e "$prefix" ] ||
 118                 die "'$prefix' does not exist; use 'git subtree add'" ;;
 119 esac
 120
 121 dir="$(dirname "$prefix/.")"
 122
 123 if [ "$command" != "pull" -a "$command" != "add" -a "$command" != "push" ]; then
 124         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 125         dirs="$(git rev-parse --no-revs --no-flags "$@")" || exit $?
 126         if [ -n "$dirs" ]; then
 127                 die "Error: Use --prefix instead of bare filenames."
 128         fi
 129 fi
 130
 131 debug "command: {$command}"
 132 debug "quiet: {$quiet}"
 133 debug "revs: {$revs}"
 134 debug "dir: {$dir}"
 135 debug "opts: {$*}"
 136 debug
 137
 138 cache_setup()
 139 {
 140         cachedir="$GIT_DIR/subtree-cache/$$"
 141         rm -rf "$cachedir" || die "Can't delete old cachedir: $cachedir"
 142         mkdir -p "$cachedir" || die "Can't create new cachedir: $cachedir"
 143         mkdir -p "$cachedir/notree" || die "Can't create new cachedir: $cachedir/notree"
 144         debug "Using cachedir: $cachedir" >&2
 145 }
 146
 147 cache_get()
 148 {
 149         for oldrev in $*; do
 150                 if [ -r "$cachedir/$oldrev" ]; then
 151                         read newrev <"$cachedir/$oldrev"
 152                         echo $newrev
 153                 fi
 154         done
 155 }
 156
 157 cache_miss()
 158 {
 159         for oldrev in $*; do
 160                 if [ ! -r "$cachedir/$oldrev" ]; then
 161                         echo $oldrev
 162                 fi
 163         done
 164 }
 165
 166 check_parents()
 167 {
 168         missed=$(cache_miss $*)
 169         for miss in $missed; do
 170                 if [ ! -r "$cachedir/notree/$miss" ]; then
 171                         debug "  incorrect order: $miss"
 172                 fi
 173         done
 174 }
 175
 176 set_notree()
 177 {
 178         echo "1" > "$cachedir/notree/$1"
 179 }
 180
 181 cache_set()
 182 {
 183         oldrev="$1"
 184         newrev="$2"
 185         if [ "$oldrev" != "latest_old" \
 186              -a "$oldrev" != "latest_new" \
 187              -a -e "$cachedir/$oldrev" ]; then
 188                 die "cache for $oldrev already exists!"
 189         fi
 190         echo "$newrev" >"$cachedir/$oldrev"
 191 }
 192
 193 rev_exists()
 194 {
 195         if git rev-parse "$1" >/dev/null 2>&1; then
 196                 return 0
 197         else
 198                 return 1
 199         fi
 200 }
 201
 202 rev_is_descendant_of_branch()
 203 {
 204         newrev="$1"
 205         branch="$2"
 206         branch_hash=$(git rev-parse $branch)
 207         match=$(git rev-list -1 $branch_hash ^$newrev)
 208
 209         if [ -z "$match" ]; then
 210                 return 0
 211         else
 212                 return 1
 213         fi
 214 }
 215
 216 # if a commit doesn't have a parent, this might not work.  But we only want
 217 # to remove the parent from the rev-list, and since it doesn't exist, it won't
 218 # be there anyway, so do nothing in that case.
 219 try_remove_previous()
 220 {
 221         if rev_exists "$1^"; then
 222                 echo "^$1^"
 223         fi
 224 }
 225
 226 find_latest_squash()
 227 {
 228         debug "Looking for latest squash ($dir)..."
 229         dir="$1"
 230         sq=
 231         main=
 232         sub=
 233         git log --grep="^git-subtree-dir: $dir/*\$" \
 234                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' HEAD |
 235         while read a b junk; do
 236                 debug "$a $b $junk"
 237                 debug "{{$sq/$main/$sub}}"
 238                 case "$a" in
 239                         START) sq="$b" ;;
 240                         git-subtree-mainline:) main="$b" ;;
 241                         git-subtree-split:) sub="$b" ;;
 242                         END)
 243                                 if [ -n "$sub" ]; then
 244                                         if [ -n "$main" ]; then
 245                                                 # a rejoin commit?
 246                                                 # Pretend its sub was a squash.
 247                                                 sq="$sub"
 248                                         fi
 249                                         debug "Squash found: $sq $sub"
 250                                         echo "$sq" "$sub"
 251                                         break
 252                                 fi
 253                                 sq=
 254                                 main=
 255                                 sub=
 256                                 ;;
 257                 esac
 258         done
 259 }
 260
 261 find_existing_splits()
 262 {
 263         debug "Looking for prior splits..."
 264         dir="$1"
 265         revs="$2"
 266         main=
 267         sub=
 268         git log --grep="^git-subtree-dir: $dir/*\$" \
 269                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' $revs |
 270         while read a b junk; do
 271                 case "$a" in
 272                         START) sq="$b" ;;
 273                         git-subtree-mainline:) main="$b" ;;
 274                         git-subtree-split:) sub="$b" ;;
 275                         END)
 276                                 debug "  Main is: '$main'"
 277                                 if [ -z "$main" -a -n "$sub" ]; then
 278                                         # squash commits refer to a subtree
 279                                         debug "  Squash: $sq from $sub"
 280                                         cache_set "$sq" "$sub"
 281                                 fi
 282                                 if [ -n "$main" -a -n "$sub" ]; then
 283                                         debug "  Prior: $main -> $sub"
 284                                         cache_set $main $sub
 285                                         cache_set $sub $sub
 286                                         try_remove_previous "$main"
 287                                         try_remove_previous "$sub"
 288                                 fi
 289                                 main=
 290                                 sub=
 291                                 ;;
 292                 esac
 293         done
 294 }
 295
 296 copy_commit()
 297 {
 298         # We're going to set some environment vars here, so
 299         # do it in a subshell to get rid of them safely later
 300         debug copy_commit "{$1}" "{$2}" "{$3}"
 301         git log -1 --pretty=format:'%an%n%ae%n%ad%n%cn%n%ce%n%cd%n%B' "$1" |
 302         (
 303                 read GIT_AUTHOR_NAME
 304                 read GIT_AUTHOR_EMAIL
 305                 read GIT_AUTHOR_DATE
 306                 read GIT_COMMITTER_NAME
 307                 read GIT_COMMITTER_EMAIL
 308                 read GIT_COMMITTER_DATE
 309                 export  GIT_AUTHOR_NAME \
 310                         GIT_AUTHOR_EMAIL \
 311                         GIT_AUTHOR_DATE \
 312                         GIT_COMMITTER_NAME \
 313                         GIT_COMMITTER_EMAIL \
 314                         GIT_COMMITTER_DATE
 315                 (printf "%s" "$annotate"; cat ) |
 316                 git commit-tree "$2" $3  # reads the rest of stdin
 317         ) || die "Can't copy commit $1"
 318 }
 319
 320 add_msg()
 321 {
 322         dir="$1"
 323         latest_old="$2"
 324         latest_new="$3"
 325         if [ -n "$message" ]; then
 326                 commit_message="$message"
 327         else
 328                 commit_message="Add '$dir/' from commit '$latest_new'"
 329         fi
 330         cat <<-EOF
 331                 $commit_message
 332
 333                 git-subtree-dir: $dir
 334                 git-subtree-mainline: $latest_old
 335                 git-subtree-split: $latest_new
 336         EOF
 337 }
 338
 339 add_squashed_msg()
 340 {
 341         if [ -n "$message" ]; then
 342                 echo "$message"
 343         else
 344                 echo "Merge commit '$1' as '$2'"
 345         fi
 346 }
 347
 348 rejoin_msg()
 349 {
 350         dir="$1"
 351         latest_old="$2"
 352         latest_new="$3"
 353         if [ -n "$message" ]; then
 354                 commit_message="$message"
 355         else
 356                 commit_message="Split '$dir/' into commit '$latest_new'"
 357         fi
 358         cat <<-EOF
 359                 $commit_message
 360
 361                 git-subtree-dir: $dir
 362                 git-subtree-mainline: $latest_old
 363                 git-subtree-split: $latest_new
 364         EOF
 365 }
 366
 367 squash_msg()
 368 {
 369         dir="$1"
 370         oldsub="$2"
 371         newsub="$3"
 372         newsub_short=$(git rev-parse --short "$newsub")
 373
 374         if [ -n "$oldsub" ]; then
 375                 oldsub_short=$(git rev-parse --short "$oldsub")
 376                 echo "Squashed '$dir/' changes from $oldsub_short..$newsub_short"
 377                 echo
 378                 git log --pretty=tformat:'%h %s' "$oldsub..$newsub"
 379                 git log --pretty=tformat:'REVERT: %h %s' "$newsub..$oldsub"
 380         else
 381                 echo "Squashed '$dir/' content from commit $newsub_short"
 382         fi
 383
 384         echo
 385         echo "git-subtree-dir: $dir"
 386         echo "git-subtree-split: $newsub"
 387 }
 388
 389 toptree_for_commit()
 390 {
 391         commit="$1"
 392         git log -1 --pretty=format:'%T' "$commit" -- || exit $?
 393 }
 394
 395 subtree_for_commit()
 396 {
 397         commit="$1"
 398         dir="$2"
 399         git ls-tree "$commit" -- "$dir" |
 400         while read mode type tree name; do
 401                 assert [ "$name" = "$dir" ]
 402                 assert [ "$type" = "tree" -o "$type" = "commit" ]
 403                 [ "$type" = "commit" ] && continue  # ignore submodules
 404                 echo $tree
 405                 break
 406         done
 407 }
 408
 409 tree_changed()
 410 {
 411         tree=$1
 412         shift
 413         if [ $# -ne 1 ]; then
 414                 return 0   # weird parents, consider it changed
 415         else
 416                 ptree=$(toptree_for_commit $1)
 417                 if [ "$ptree" != "$tree" ]; then
 418                         return 0   # changed
 419                 else
 420                         return 1   # not changed
 421                 fi
 422         fi
 423 }
 424
 425 new_squash_commit()
 426 {
 427         old="$1"
 428         oldsub="$2"
 429         newsub="$3"
 430         tree=$(toptree_for_commit $newsub) || exit $?
 431         if [ -n "$old" ]; then
 432                 squash_msg "$dir" "$oldsub" "$newsub" |
 433                         git commit-tree "$tree" -p "$old" || exit $?
 434         else
 435                 squash_msg "$dir" "" "$newsub" |
 436                         git commit-tree "$tree" || exit $?
 437         fi
 438 }
 439
 440 copy_or_skip()
 441 {
 442         rev="$1"
 443         tree="$2"
 444         newparents="$3"
 445         assert [ -n "$tree" ]
 446
 447         identical=
 448         nonidentical=
 449         p=
 450         gotparents=
 451         for parent in $newparents; do
 452                 ptree=$(toptree_for_commit $parent) || exit $?
 453                 [ -z "$ptree" ] && continue
 454                 if [ "$ptree" = "$tree" ]; then
 455                         # an identical parent could be used in place of this rev.
 456                         identical="$parent"
 457                 else
 458                         nonidentical="$parent"
 459                 fi
 460
 461                 # sometimes both old parents map to the same newparent;
 462                 # eliminate duplicates
 463                 is_new=1
 464                 for gp in $gotparents; do
 465                         if [ "$gp" = "$parent" ]; then
 466                                 is_new=
 467                                 break
 468                         fi
 469                 done
 470                 if [ -n "$is_new" ]; then
 471                         gotparents="$gotparents $parent"
 472                         p="$p -p $parent"
 473                 fi
 474         done
 475
 476         if [ -n "$identical" ]; then
 477                 echo $identical
 478         else
 479                 copy_commit $rev $tree "$p" || exit $?
 480         fi
 481 }
 482
 483 ensure_clean()
 484 {
 485         if ! git diff-index HEAD --exit-code --quiet 2>&1; then
 486                 die "Working tree has modifications.  Cannot add."
 487         fi
 488         if ! git diff-index --cached HEAD --exit-code --quiet 2>&1; then
 489                 die "Index has modifications.  Cannot add."
 490         fi
 491 }
 492
 493 ensure_valid_ref_format()
 494 {
 495         git check-ref-format "refs/heads/$1" ||
 496             die "'$1' does not look like a ref"
 497 }
 498
 499 cmd_add()
 500 {
 501         if [ -e "$dir" ]; then
 502                 die "'$dir' already exists.  Cannot add."
 503         fi
 504
 505         ensure_clean
 506
 507         if [ $# -eq 1 ]; then
 508             git rev-parse -q --verify "$1^{commit}" >/dev/null ||
 509             die "'$1' does not refer to a commit"
 510
 511             "cmd_add_commit" "$@"
 512         elif [ $# -eq 2 ]; then
 513             # Technically we could accept a refspec here but we're
 514             # just going to turn around and add FETCH_HEAD under the
 515             # specified directory.  Allowing a refspec might be
 516             # misleading because we won't do anything with any other
 517             # branches fetched via the refspec.
 518             ensure_valid_ref_format "$2"
 519
 520             "cmd_add_repository" "$@"
 521         else
 522             say "error: parameters were '$@'"
 523             die "Provide either a commit or a repository and commit."
 524         fi
 525 }
 526
 527 cmd_add_repository()
 528 {
 529         echo "git fetch" "$@"
 530         repository=$1
 531         refspec=$2
 532         git fetch "$@" || exit $?
 533         revs=FETCH_HEAD
 534         set -- $revs
 535         cmd_add_commit "$@"
 536 }
 537
 538 cmd_add_commit()
 539 {
 540         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 541         set -- $revs
 542         rev="$1"
 543
 544         debug "Adding $dir as '$rev'..."
 545         git read-tree --prefix="$dir" $rev || exit $?
 546         git checkout -- "$dir" || exit $?
 547         tree=$(git write-tree) || exit $?
 548
 549         headrev=$(git rev-parse HEAD) || exit $?
 550         if [ -n "$headrev" -a "$headrev" != "$rev" ]; then
 551                 headp="-p $headrev"
 552         else
 553                 headp=
 554         fi
 555
 556         if [ -n "$squash" ]; then
 557                 rev=$(new_squash_commit "" "" "$rev") || exit $?
 558                 commit=$(add_squashed_msg "$rev" "$dir" |
 559                          git commit-tree $tree $headp -p "$rev") || exit $?
 560         else
 561                 commit=$(add_msg "$dir" "$headrev" "$rev" |
 562                          git commit-tree $tree $headp -p "$rev") || exit $?
 563         fi
 564         git reset "$commit" || exit $?
 565
 566         say "Added dir '$dir'"
 567 }
 568
 569 cmd_split()
 570 {
 571         debug "Splitting $dir..."
 572         cache_setup || exit $?
 573
 574         if [ -n "$onto" ]; then
 575                 debug "Reading history for --onto=$onto..."
 576                 git rev-list $onto |
 577                 while read rev; do
 578                         # the 'onto' history is already just the subdir, so
 579                         # any parent we find there can be used verbatim
 580                         debug "  cache: $rev"
 581                         cache_set $rev $rev
 582                 done
 583         fi
 584
 585         if [ -n "$ignore_joins" ]; then
 586                 unrevs=
 587         else
 588                 unrevs="$(find_existing_splits "$dir" "$revs")"
 589         fi
 590
 591         # We can't restrict rev-list to only $dir here, because some of our
 592         # parents have the $dir contents the root, and those won't match.
 593         # (and rev-list --follow doesn't seem to solve this)
 594         grl='git rev-list --topo-order --reverse --parents $revs $unrevs'
 595         revmax=$(eval "$grl" | wc -l)
 596         revcount=0
 597         createcount=0
 598         eval "$grl" |
 599         while read rev parents; do
 600                 revcount=$(($revcount + 1))
 601                 say -n "$revcount/$revmax ($createcount)
"
 602                 debug "Processing commit: $rev"
 603                 exists=$(cache_get $rev)
 604                 if [ -n "$exists" ]; then
 605                         debug "  prior: $exists"
 606                         continue
 607                 fi
 608                 createcount=$(($createcount + 1))
 609                 debug "  parents: $parents"
 610                 newparents=$(cache_get $parents)
 611                 debug "  newparents: $newparents"
 612
 613                 tree=$(subtree_for_commit $rev "$dir")
 614                 debug "  tree is: $tree"
 615
 616                 check_parents $parents
 617
 618                 # ugly.  is there no better way to tell if this is a subtree
 619                 # vs. a mainline commit?  Does it matter?
 620                 if [ -z $tree ]; then
 621                         set_notree $rev
 622                         if [ -n "$newparents" ]; then
 623                                 cache_set $rev $rev
 624                         fi
 625                         continue
 626                 fi
 627
 628                 newrev=$(copy_or_skip "$rev" "$tree" "$newparents") || exit $?
 629                 debug "  newrev is: $newrev"
 630                 cache_set $rev $newrev
 631                 cache_set latest_new $newrev
 632                 cache_set latest_old $rev
 633         done || exit $?
 634         latest_new=$(cache_get latest_new)
 635         if [ -z "$latest_new" ]; then
 636                 die "No new revisions were found"
 637         fi
 638
 639         if [ -n "$rejoin" ]; then
 640                 debug "Merging split branch into HEAD..."
 641                 latest_old=$(cache_get latest_old)
 642                 git merge -s ours \
 643                         -m "$(rejoin_msg $dir $latest_old $latest_new)" \
 644                         $latest_new >&2 || exit $?
 645         fi
 646         if [ -n "$branch" ]; then
 647                 if rev_exists "refs/heads/$branch"; then
 648                         if ! rev_is_descendant_of_branch $latest_new $branch; then
 649                                 die "Branch '$branch' is not an ancestor of commit '$latest_new'."
 650                         fi
 651                         action='Updated'
 652                 else
 653                         action='Created'
 654                 fi
 655                 git update-ref -m 'subtree split' "refs/heads/$branch" $latest_new || exit $?
 656                 say "$action branch '$branch'"
 657         fi
 658         echo $latest_new
 659         exit 0
 660 }
 661
 662 cmd_merge()
 663 {
 664         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 665         ensure_clean
 666
 667         set -- $revs
 668         if [ $# -ne 1 ]; then
 669                 die "You must provide exactly one revision.  Got: '$revs'"
 670         fi
 671         rev="$1"
 672
 673         if [ -n "$squash" ]; then
 674                 first_split="$(find_latest_squash "$dir")"
 675                 if [ -z "$first_split" ]; then
 676                         die "Can't squash-merge: '$dir' was never added."
 677                 fi
 678                 set $first_split
 679                 old=$1
 680                 sub=$2
 681                 if [ "$sub" = "$rev" ]; then
 682                         say "Subtree is already at commit $rev."
 683                         exit 0
 684                 fi
 685                 new=$(new_squash_commit "$old" "$sub" "$rev") || exit $?
 686                 debug "New squash commit: $new"
 687                 rev="$new"
 688         fi
 689
 690         version=$(git version)
 691         if [ "$version" \< "git version 1.7" ]; then
 692                 if [ -n "$message" ]; then
 693                         git merge -s subtree --message="$message" $rev
 694                 else
 695                         git merge -s subtree $rev
 696                 fi
 697         else
 698                 if [ -n "$message" ]; then
 699                         git merge -Xsubtree="$prefix" --message="$message" $rev
 700                 else
 701                         git merge -Xsubtree="$prefix" $rev
 702                 fi
 703         fi
 704 }
 705
 706 cmd_pull()
 707 {
 708         if [ $# -ne 2 ]; then
 709             die "You must provide <repository> <ref>"
 710         fi
 711         ensure_clean
 712         ensure_valid_ref_format "$2"
 713         git fetch "$@" || exit $?
 714         revs=FETCH_HEAD
 715         set -- $revs
 716         cmd_merge "$@"
 717 }
 718
 719 cmd_push()
 720 {
 721         if [ $# -ne 2 ]; then
 722             die "You must provide <repository> <ref>"
 723         fi
 724         ensure_valid_ref_format "$2"
 725         if [ -e "$dir" ]; then
 726             repository=$1
 727             refspec=$2
 728             echo "git push using: " $repository $refspec
 729             localrev=$(git subtree split --prefix="$prefix") || die
 730             git push $repository $localrev:refs/heads/$refspec
 731         else
 732             die "'$dir' must already exist. Try 'git subtree add'."
 733         fi
 734 }
 735
 736 "cmd_$command" "$@"