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 [ $# -eq 0 ]; then
   8     set -- -h
   9 fi
  10 OPTS_SPEC="\
  11 git subtree add   --prefix=<prefix> <commit>
  12 git subtree add   --prefix=<prefix> <repository> <ref>
  13 git subtree merge --prefix=<prefix> <commit>
  14 git subtree pull  --prefix=<prefix> <repository> <ref>
  15 git subtree push  --prefix=<prefix> <repository> <ref>
  16 git subtree split --prefix=<prefix> <commit...>
  17 --
  18 h,help        show the help
  19 q             quiet
  20 d             show debug messages
  21 P,prefix=     the name of the subdir to split out
  22 m,message=    use the given message as the commit message for the merge commit
  23  options for 'split'
  24 annotate=     add a prefix to commit message of new commits
  25 b,branch=     create a new branch from the split subtree
  26 ignore-joins  ignore prior --rejoin commits
  27 onto=         try connecting new tree to an existing one
  28 rejoin        merge the new branch back into HEAD
  29  options for 'add', 'merge', and 'pull'
  30 squash        merge subtree changes as a single commit
  31 "
  32 eval "$(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                 printf "%s\n" "$*" >&2
  55         fi
  56 }
  57
  58 say()
  59 {
  60         if [ -z "$quiet" ]; then
  61                 printf "%s\n" "$*" >&2
  62         fi
  63 }
  64
  65 progress()
  66 {
  67         if [ -z "$quiet" ]; then
  68                 printf "%s\r" "$*" >&2
  69         fi
  70 }
  71
  72 assert()
  73 {
  74         if "$@"; then
  75                 :
  76         else
  77                 die "assertion failed: " "$@"
  78         fi
  79 }
  80
  81
  82 #echo "Options: $*"
  83
  84 while [ $# -gt 0 ]; do
  85         opt="$1"
  86         shift
  87         case "$opt" in
  88                 -q) quiet=1 ;;
  89                 -d) debug=1 ;;
  90                 --annotate) annotate="$1"; shift ;;
  91                 --no-annotate) annotate= ;;
  92                 -b) branch="$1"; shift ;;
  93                 -P) prefix="${1%/}"; shift ;;
  94                 -m) message="$1"; shift ;;
  95                 --no-prefix) prefix= ;;
  96                 --onto) onto="$1"; shift ;;
  97                 --no-onto) onto= ;;
  98                 --rejoin) rejoin=1 ;;
  99                 --no-rejoin) rejoin= ;;
 100                 --ignore-joins) ignore_joins=1 ;;
 101                 --no-ignore-joins) ignore_joins= ;;
 102                 --squash) squash=1 ;;
 103                 --no-squash) squash= ;;
 104                 --) break ;;
 105                 *) die "Unexpected option: $opt" ;;
 106         esac
 107 done
 108
 109 command="$1"
 110 shift
 111 case "$command" in
 112         add|merge|pull) default= ;;
 113         split|push) default="--default HEAD" ;;
 114         *) die "Unknown command '$command'" ;;
 115 esac
 116
 117 if [ -z "$prefix" ]; then
 118         die "You must provide the --prefix option."
 119 fi
 120
 121 case "$command" in
 122         add) [ -e "$prefix" ] &&
 123                 die "prefix '$prefix' already exists." ;;
 124         *)   [ -e "$prefix" ] ||
 125                 die "'$prefix' does not exist; use 'git subtree add'" ;;
 126 esac
 127
 128 dir="$(dirname "$prefix/.")"
 129
 130 if [ "$command" != "pull" -a "$command" != "add" -a "$command" != "push" ]; then
 131         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 132         dirs="$(git rev-parse --no-revs --no-flags "$@")" || exit $?
 133         if [ -n "$dirs" ]; then
 134                 die "Error: Use --prefix instead of bare filenames."
 135         fi
 136 fi
 137
 138 debug "command: {$command}"
 139 debug "quiet: {$quiet}"
 140 debug "revs: {$revs}"
 141 debug "dir: {$dir}"
 142 debug "opts: {$*}"
 143 debug
 144
 145 cache_setup()
 146 {
 147         cachedir="$GIT_DIR/subtree-cache/$$"
 148         rm -rf "$cachedir" || die "Can't delete old cachedir: $cachedir"
 149         mkdir -p "$cachedir" || die "Can't create new cachedir: $cachedir"
 150         mkdir -p "$cachedir/notree" || die "Can't create new cachedir: $cachedir/notree"
 151         debug "Using cachedir: $cachedir" >&2
 152 }
 153
 154 cache_get()
 155 {
 156         for oldrev in $*; do
 157                 if [ -r "$cachedir/$oldrev" ]; then
 158                         read newrev <"$cachedir/$oldrev"
 159                         echo $newrev
 160                 fi
 161         done
 162 }
 163
 164 cache_miss()
 165 {
 166         for oldrev in $*; do
 167                 if [ ! -r "$cachedir/$oldrev" ]; then
 168                         echo $oldrev
 169                 fi
 170         done
 171 }
 172
 173 check_parents()
 174 {
 175         missed=$(cache_miss $*)
 176         for miss in $missed; do
 177                 if [ ! -r "$cachedir/notree/$miss" ]; then
 178                         debug "  incorrect order: $miss"
 179                 fi
 180         done
 181 }
 182
 183 set_notree()
 184 {
 185         echo "1" > "$cachedir/notree/$1"
 186 }
 187
 188 cache_set()
 189 {
 190         oldrev="$1"
 191         newrev="$2"
 192         if [ "$oldrev" != "latest_old" \
 193              -a "$oldrev" != "latest_new" \
 194              -a -e "$cachedir/$oldrev" ]; then
 195                 die "cache for $oldrev already exists!"
 196         fi
 197         echo "$newrev" >"$cachedir/$oldrev"
 198 }
 199
 200 rev_exists()
 201 {
 202         if git rev-parse "$1" >/dev/null 2>&1; then
 203                 return 0
 204         else
 205                 return 1
 206         fi
 207 }
 208
 209 rev_is_descendant_of_branch()
 210 {
 211         newrev="$1"
 212         branch="$2"
 213         branch_hash=$(git rev-parse $branch)
 214         match=$(git rev-list -1 $branch_hash ^$newrev)
 215
 216         if [ -z "$match" ]; then
 217                 return 0
 218         else
 219                 return 1
 220         fi
 221 }
 222
 223 # if a commit doesn't have a parent, this might not work.  But we only want
 224 # to remove the parent from the rev-list, and since it doesn't exist, it won't
 225 # be there anyway, so do nothing in that case.
 226 try_remove_previous()
 227 {
 228         if rev_exists "$1^"; then
 229                 echo "^$1^"
 230         fi
 231 }
 232
 233 find_latest_squash()
 234 {
 235         debug "Looking for latest squash ($dir)..."
 236         dir="$1"
 237         sq=
 238         main=
 239         sub=
 240         git log --grep="^git-subtree-dir: $dir/*\$" \
 241                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' HEAD |
 242         while read a b junk; do
 243                 debug "$a $b $junk"
 244                 debug "{{$sq/$main/$sub}}"
 245                 case "$a" in
 246                         START) sq="$b" ;;
 247                         git-subtree-mainline:) main="$b" ;;
 248                         git-subtree-split:) sub="$b" ;;
 249                         END)
 250                                 if [ -n "$sub" ]; then
 251                                         if [ -n "$main" ]; then
 252                                                 # a rejoin commit?
 253                                                 # Pretend its sub was a squash.
 254                                                 sq="$sub"
 255                                         fi
 256                                         debug "Squash found: $sq $sub"
 257                                         echo "$sq" "$sub"
 258                                         break
 259                                 fi
 260                                 sq=
 261                                 main=
 262                                 sub=
 263                                 ;;
 264                 esac
 265         done
 266 }
 267
 268 find_existing_splits()
 269 {
 270         debug "Looking for prior splits..."
 271         dir="$1"
 272         revs="$2"
 273         main=
 274         sub=
 275         git log --grep="^git-subtree-dir: $dir/*\$" \
 276                 --pretty=format:'START %H%n%s%n%n%b%nEND%n' $revs |
 277         while read a b junk; do
 278                 case "$a" in
 279                         START) sq="$b" ;;
 280                         git-subtree-mainline:) main="$b" ;;
 281                         git-subtree-split:) sub="$b" ;;
 282                         END)
 283                                 debug "  Main is: '$main'"
 284                                 if [ -z "$main" -a -n "$sub" ]; then
 285                                         # squash commits refer to a subtree
 286                                         debug "  Squash: $sq from $sub"
 287                                         cache_set "$sq" "$sub"
 288                                 fi
 289                                 if [ -n "$main" -a -n "$sub" ]; then
 290                                         debug "  Prior: $main -> $sub"
 291                                         cache_set $main $sub
 292                                         cache_set $sub $sub
 293                                         try_remove_previous "$main"
 294                                         try_remove_previous "$sub"
 295                                 fi
 296                                 main=
 297                                 sub=
 298                                 ;;
 299                 esac
 300         done
 301 }
 302
 303 copy_commit()
 304 {
 305         # We're going to set some environment vars here, so
 306         # do it in a subshell to get rid of them safely later
 307         debug copy_commit "{$1}" "{$2}" "{$3}"
 308         git log -1 --pretty=format:'%an%n%ae%n%aD%n%cn%n%ce%n%cD%n%B' "$1" |
 309         (
 310                 read GIT_AUTHOR_NAME
 311                 read GIT_AUTHOR_EMAIL
 312                 read GIT_AUTHOR_DATE
 313                 read GIT_COMMITTER_NAME
 314                 read GIT_COMMITTER_EMAIL
 315                 read GIT_COMMITTER_DATE
 316                 export  GIT_AUTHOR_NAME \
 317                         GIT_AUTHOR_EMAIL \
 318                         GIT_AUTHOR_DATE \
 319                         GIT_COMMITTER_NAME \
 320                         GIT_COMMITTER_EMAIL \
 321                         GIT_COMMITTER_DATE
 322                 (printf "%s" "$annotate"; cat ) |
 323                 git commit-tree "$2" $3  # reads the rest of stdin
 324         ) || die "Can't copy commit $1"
 325 }
 326
 327 add_msg()
 328 {
 329         dir="$1"
 330         latest_old="$2"
 331         latest_new="$3"
 332         if [ -n "$message" ]; then
 333                 commit_message="$message"
 334         else
 335                 commit_message="Add '$dir/' from commit '$latest_new'"
 336         fi
 337         cat <<-EOF
 338                 $commit_message
 339
 340                 git-subtree-dir: $dir
 341                 git-subtree-mainline: $latest_old
 342                 git-subtree-split: $latest_new
 343         EOF
 344 }
 345
 346 add_squashed_msg()
 347 {
 348         if [ -n "$message" ]; then
 349                 echo "$message"
 350         else
 351                 echo "Merge commit '$1' as '$2'"
 352         fi
 353 }
 354
 355 rejoin_msg()
 356 {
 357         dir="$1"
 358         latest_old="$2"
 359         latest_new="$3"
 360         if [ -n "$message" ]; then
 361                 commit_message="$message"
 362         else
 363                 commit_message="Split '$dir/' into commit '$latest_new'"
 364         fi
 365         cat <<-EOF
 366                 $commit_message
 367
 368                 git-subtree-dir: $dir
 369                 git-subtree-mainline: $latest_old
 370                 git-subtree-split: $latest_new
 371         EOF
 372 }
 373
 374 squash_msg()
 375 {
 376         dir="$1"
 377         oldsub="$2"
 378         newsub="$3"
 379         newsub_short=$(git rev-parse --short "$newsub")
 380
 381         if [ -n "$oldsub" ]; then
 382                 oldsub_short=$(git rev-parse --short "$oldsub")
 383                 echo "Squashed '$dir/' changes from $oldsub_short..$newsub_short"
 384                 echo
 385                 git log --pretty=tformat:'%h %s' "$oldsub..$newsub"
 386                 git log --pretty=tformat:'REVERT: %h %s' "$newsub..$oldsub"
 387         else
 388                 echo "Squashed '$dir/' content from commit $newsub_short"
 389         fi
 390
 391         echo
 392         echo "git-subtree-dir: $dir"
 393         echo "git-subtree-split: $newsub"
 394 }
 395
 396 toptree_for_commit()
 397 {
 398         commit="$1"
 399         git log -1 --pretty=format:'%T' "$commit" -- || exit $?
 400 }
 401
 402 subtree_for_commit()
 403 {
 404         commit="$1"
 405         dir="$2"
 406         git ls-tree "$commit" -- "$dir" |
 407         while read mode type tree name; do
 408                 assert [ "$name" = "$dir" ]
 409                 assert [ "$type" = "tree" -o "$type" = "commit" ]
 410                 [ "$type" = "commit" ] && continue  # ignore submodules
 411                 echo $tree
 412                 break
 413         done
 414 }
 415
 416 tree_changed()
 417 {
 418         tree=$1
 419         shift
 420         if [ $# -ne 1 ]; then
 421                 return 0   # weird parents, consider it changed
 422         else
 423                 ptree=$(toptree_for_commit $1)
 424                 if [ "$ptree" != "$tree" ]; then
 425                         return 0   # changed
 426                 else
 427                         return 1   # not changed
 428                 fi
 429         fi
 430 }
 431
 432 new_squash_commit()
 433 {
 434         old="$1"
 435         oldsub="$2"
 436         newsub="$3"
 437         tree=$(toptree_for_commit $newsub) || exit $?
 438         if [ -n "$old" ]; then
 439                 squash_msg "$dir" "$oldsub" "$newsub" |
 440                         git commit-tree "$tree" -p "$old" || exit $?
 441         else
 442                 squash_msg "$dir" "" "$newsub" |
 443                         git commit-tree "$tree" || exit $?
 444         fi
 445 }
 446
 447 copy_or_skip()
 448 {
 449         rev="$1"
 450         tree="$2"
 451         newparents="$3"
 452         assert [ -n "$tree" ]
 453
 454         identical=
 455         nonidentical=
 456         p=
 457         gotparents=
 458         for parent in $newparents; do
 459                 ptree=$(toptree_for_commit $parent) || exit $?
 460                 [ -z "$ptree" ] && continue
 461                 if [ "$ptree" = "$tree" ]; then
 462                         # an identical parent could be used in place of this rev.
 463                         identical="$parent"
 464                 else
 465                         nonidentical="$parent"
 466                 fi
 467
 468                 # sometimes both old parents map to the same newparent;
 469                 # eliminate duplicates
 470                 is_new=1
 471                 for gp in $gotparents; do
 472                         if [ "$gp" = "$parent" ]; then
 473                                 is_new=
 474                                 break
 475                         fi
 476                 done
 477                 if [ -n "$is_new" ]; then
 478                         gotparents="$gotparents $parent"
 479                         p="$p -p $parent"
 480                 fi
 481         done
 482
 483         if [ -n "$identical" ]; then
 484                 echo $identical
 485         else
 486                 copy_commit $rev $tree "$p" || exit $?
 487         fi
 488 }
 489
 490 ensure_clean()
 491 {
 492         if ! git diff-index HEAD --exit-code --quiet 2>&1; then
 493                 die "Working tree has modifications.  Cannot add."
 494         fi
 495         if ! git diff-index --cached HEAD --exit-code --quiet 2>&1; then
 496                 die "Index has modifications.  Cannot add."
 497         fi
 498 }
 499
 500 ensure_valid_ref_format()
 501 {
 502         git check-ref-format "refs/heads/$1" ||
 503             die "'$1' does not look like a ref"
 504 }
 505
 506 cmd_add()
 507 {
 508         if [ -e "$dir" ]; then
 509                 die "'$dir' already exists.  Cannot add."
 510         fi
 511
 512         ensure_clean
 513
 514         if [ $# -eq 1 ]; then
 515             git rev-parse -q --verify "$1^{commit}" >/dev/null ||
 516             die "'$1' does not refer to a commit"
 517
 518             "cmd_add_commit" "$@"
 519         elif [ $# -eq 2 ]; then
 520             # Technically we could accept a refspec here but we're
 521             # just going to turn around and add FETCH_HEAD under the
 522             # specified directory.  Allowing a refspec might be
 523             # misleading because we won't do anything with any other
 524             # branches fetched via the refspec.
 525             ensure_valid_ref_format "$2"
 526
 527             "cmd_add_repository" "$@"
 528         else
 529             say "error: parameters were '$@'"
 530             die "Provide either a commit or a repository and commit."
 531         fi
 532 }
 533
 534 cmd_add_repository()
 535 {
 536         echo "git fetch" "$@"
 537         repository=$1
 538         refspec=$2
 539         git fetch "$@" || exit $?
 540         revs=FETCH_HEAD
 541         set -- $revs
 542         cmd_add_commit "$@"
 543 }
 544
 545 cmd_add_commit()
 546 {
 547         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 548         set -- $revs
 549         rev="$1"
 550
 551         debug "Adding $dir as '$rev'..."
 552         git read-tree --prefix="$dir" $rev || exit $?
 553         git checkout -- "$dir" || exit $?
 554         tree=$(git write-tree) || exit $?
 555
 556         headrev=$(git rev-parse HEAD) || exit $?
 557         if [ -n "$headrev" -a "$headrev" != "$rev" ]; then
 558                 headp="-p $headrev"
 559         else
 560                 headp=
 561         fi
 562
 563         if [ -n "$squash" ]; then
 564                 rev=$(new_squash_commit "" "" "$rev") || exit $?
 565                 commit=$(add_squashed_msg "$rev" "$dir" |
 566                          git commit-tree $tree $headp -p "$rev") || exit $?
 567         else
 568                 revp=$(peel_committish "$rev") &&
 569                 commit=$(add_msg "$dir" "$headrev" "$rev" |
 570                          git commit-tree $tree $headp -p "$revp") || exit $?
 571         fi
 572         git reset "$commit" || exit $?
 573
 574         say "Added dir '$dir'"
 575 }
 576
 577 cmd_split()
 578 {
 579         debug "Splitting $dir..."
 580         cache_setup || exit $?
 581
 582         if [ -n "$onto" ]; then
 583                 debug "Reading history for --onto=$onto..."
 584                 git rev-list $onto |
 585                 while read rev; do
 586                         # the 'onto' history is already just the subdir, so
 587                         # any parent we find there can be used verbatim
 588                         debug "  cache: $rev"
 589                         cache_set $rev $rev
 590                 done
 591         fi
 592
 593         if [ -n "$ignore_joins" ]; then
 594                 unrevs=
 595         else
 596                 unrevs="$(find_existing_splits "$dir" "$revs")"
 597         fi
 598
 599         # We can't restrict rev-list to only $dir here, because some of our
 600         # parents have the $dir contents the root, and those won't match.
 601         # (and rev-list --follow doesn't seem to solve this)
 602         grl='git rev-list --topo-order --reverse --parents $revs $unrevs'
 603         revmax=$(eval "$grl" | wc -l)
 604         revcount=0
 605         createcount=0
 606         eval "$grl" |
 607         while read rev parents; do
 608                 revcount=$(($revcount + 1))
 609                 progress "$revcount/$revmax ($createcount)"
 610                 debug "Processing commit: $rev"
 611                 exists=$(cache_get $rev)
 612                 if [ -n "$exists" ]; then
 613                         debug "  prior: $exists"
 614                         continue
 615                 fi
 616                 createcount=$(($createcount + 1))
 617                 debug "  parents: $parents"
 618                 newparents=$(cache_get $parents)
 619                 debug "  newparents: $newparents"
 620
 621                 tree=$(subtree_for_commit $rev "$dir")
 622                 debug "  tree is: $tree"
 623
 624                 check_parents $parents
 625
 626                 # ugly.  is there no better way to tell if this is a subtree
 627                 # vs. a mainline commit?  Does it matter?
 628                 if [ -z $tree ]; then
 629                         set_notree $rev
 630                         if [ -n "$newparents" ]; then
 631                                 cache_set $rev $rev
 632                         fi
 633                         continue
 634                 fi
 635
 636                 newrev=$(copy_or_skip "$rev" "$tree" "$newparents") || exit $?
 637                 debug "  newrev is: $newrev"
 638                 cache_set $rev $newrev
 639                 cache_set latest_new $newrev
 640                 cache_set latest_old $rev
 641         done || exit $?
 642         latest_new=$(cache_get latest_new)
 643         if [ -z "$latest_new" ]; then
 644                 die "No new revisions were found"
 645         fi
 646
 647         if [ -n "$rejoin" ]; then
 648                 debug "Merging split branch into HEAD..."
 649                 latest_old=$(cache_get latest_old)
 650                 git merge -s ours \
 651                         -m "$(rejoin_msg "$dir" $latest_old $latest_new)" \
 652                         $latest_new >&2 || exit $?
 653         fi
 654         if [ -n "$branch" ]; then
 655                 if rev_exists "refs/heads/$branch"; then
 656                         if ! rev_is_descendant_of_branch $latest_new $branch; then
 657                                 die "Branch '$branch' is not an ancestor of commit '$latest_new'."
 658                         fi
 659                         action='Updated'
 660                 else
 661                         action='Created'
 662                 fi
 663                 git update-ref -m 'subtree split' "refs/heads/$branch" $latest_new || exit $?
 664                 say "$action branch '$branch'"
 665         fi
 666         echo $latest_new
 667         exit 0
 668 }
 669
 670 cmd_merge()
 671 {
 672         revs=$(git rev-parse $default --revs-only "$@") || exit $?
 673         ensure_clean
 674
 675         set -- $revs
 676         if [ $# -ne 1 ]; then
 677                 die "You must provide exactly one revision.  Got: '$revs'"
 678         fi
 679         rev="$1"
 680
 681         if [ -n "$squash" ]; then
 682                 first_split="$(find_latest_squash "$dir")"
 683                 if [ -z "$first_split" ]; then
 684                         die "Can't squash-merge: '$dir' was never added."
 685                 fi
 686                 set $first_split
 687                 old=$1
 688                 sub=$2
 689                 if [ "$sub" = "$rev" ]; then
 690                         say "Subtree is already at commit $rev."
 691                         exit 0
 692                 fi
 693                 new=$(new_squash_commit "$old" "$sub" "$rev") || exit $?
 694                 debug "New squash commit: $new"
 695                 rev="$new"
 696         fi
 697
 698         version=$(git version)
 699         if [ "$version" \< "git version 1.7" ]; then
 700                 if [ -n "$message" ]; then
 701                         git merge -s subtree --message="$message" $rev
 702                 else
 703                         git merge -s subtree $rev
 704                 fi
 705         else
 706                 if [ -n "$message" ]; then
 707                         git merge -Xsubtree="$prefix" --message="$message" $rev
 708                 else
 709                         git merge -Xsubtree="$prefix" $rev
 710                 fi
 711         fi
 712 }
 713
 714 cmd_pull()
 715 {
 716         if [ $# -ne 2 ]; then
 717             die "You must provide <repository> <ref>"
 718         fi
 719         ensure_clean
 720         ensure_valid_ref_format "$2"
 721         git fetch "$@" || exit $?
 722         revs=FETCH_HEAD
 723         set -- $revs
 724         cmd_merge "$@"
 725 }
 726
 727 cmd_push()
 728 {
 729         if [ $# -ne 2 ]; then
 730             die "You must provide <repository> <ref>"
 731         fi
 732         ensure_valid_ref_format "$2"
 733         if [ -e "$dir" ]; then
 734             repository=$1
 735             refspec=$2
 736             echo "git push using: " $repository $refspec
 737             localrev=$(git subtree split --prefix="$prefix") || die
 738             git push "$repository" $localrev:refs/heads/$refspec
 739         else
 740             die "'$dir' must already exist. Try 'git subtree add'."
 741         fi
 742 }
 743
 744 "cmd_$command" "$@"