svn-merge2git.sh

   1 #!/bin/sh
   2 # Copyright (c) 2008  Benoit Sigoure <tsuna@lrde.epita.fr>
   3 #
   4 # This program is free software: you can redistribute it and/or modify
   5 # it under the terms of the GNU General Public License as published by
   6 # the Free Software Foundation, either version 3 of the License, or
   7 # (at your option) any later version.
   8 #
   9 # This program is distributed in the hope that it will be useful,
  10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 # GNU General Public License for more details.
  13 #
  14 # You should have received a copy of the GNU General Public License
  15 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
  16
  17 OPTIONS_SPEC="\
  18 svn-merge2git [options] [refspec]
  19 --
  20 a,all      Do the work on all local the branches
  21 d,debug    Turn on debug mode (useful if you're hacking the script)
  22 n,dry-run  Do the entire processing without actually changing anything
  23 r,remote   Include remote branches (create a local branch for each of them)
  24 v,verbose  Be more verbose
  25 "
  26 SUBDIRECTORY_OK=Yes
  27 . git-sh-setup
  28 cd_to_toplevel
  29
  30 : ${TMPDIR=/tmp}
  31 export TMPDIR
  32
  33 # BRE (Basic RegExp) compatible with `git rev-list --grep' and `sed'.  The RE
  34 # *must* capture the revision merged in its first group.
  35 merge_pattern='[Mm]erge.*[0-9][0-9]*:\([0-9][0-9]*\)'
  36
  37 # BRE which is used to exclude commits whose line that matches of
  38 # $merge_pattern also match this pattern.
  39 exclude_pattern='Finish'
  40
  41 # BRE which is used to exclude matches in the commit log of potential merge
  42 # commits.
  43 log_exclude_pattern='Finish.*merge'
  44
  45 # extract_svn_branch_name <string>
  46 # --------------------------------
  47 # Find the string the name of a SVN branch.  Put the result in
  48 # $svn_branch_name.  Assumes SVN "stdlayout".
  49 extract_svn_branch_name()
  50 {
  51   case $1 in
  52     '')
  53       fatal 'extract_svn_branch_name called with empty argument';;
  54     */branches/*)
  55       extract_svn_branch_name_ 'branches' "$1";;
  56     */tags/*)
  57       extract_svn_branch_name_ 'tags' "$1"
  58       warn "found a merge from tag '$svn_branch_name'";;
  59     */trunk*)
  60       svn_branch_name='trunk';;
  61   esac
  62 }
  63
  64 # extract_svn_branch_name_ <kind> <string>
  65 # ----------------------------------------
  66 # Helper of extract_svn_branch_name below to factor some code.
  67 # <kind> is probably either 'branches' or 'tags' (for SVN "stdlayout").
  68 # Put the result in $svn_branch_name.
  69 extract_svn_branch_name_()
  70 {
  71   # XXX: Assumes that a branch name does contain a whitespace.  Fragile.
  72   sed_tmp="s|.*/\\($1/[^        ]*\\).*|\\1|"
  73   svn_branch_name=`echo "$2" | sed "$sed_tmp"`
  74 }
  75
  76 me=`basename "$0"`
  77
  78 # fatal <msg>
  79 # -----------
  80 # print <msg> on stderr and exit 1
  81 fatal()
  82 {
  83   die "$me: error: $*"
  84 }
  85
  86 warnings=0
  87 # warn <msg>
  88 # ----------
  89 # print <msg> on stderr
  90 warn()
  91 {
  92   echo "$me: warning: $*" >&2
  93   warnings=$(($warnings + 1))
  94 }
  95
  96 # verb <msg>
  97 # ----------
  98 # Print <msg> when verbose mode is enabled.
  99 verb()
 100 {
 101   $opt_verbose && echo "$*"
 102 }
 103
 104 # debug <msg>
 105 # -----------
 106 # Print <msg> when debug mode is enabled.
 107 debug()
 108 {
 109   $opt_debug && echo "$*"
 110 }
 111
 112 # find_merge_parent <ref> <merge-line>
 113 # ------------------------------------
 114 # Return (in $merge_parent) the sha1 of the commit that has been merged in by
 115 # <ref>.  <merge-line> must be a line extracted from the commit message of
 116 # <ref> and will be used to extract the SVN revision merged.  For instance, if
 117 # <ref> is a SVN merge of merge-line='Merge -r42:51 in branch foo', this
 118 # function puts the sha1 of the first commit the revision of which is <= 51
 119 # which happens to be in branch foo in $merge_parent.
 120 # If the name of the branch being merged couldn't be found, $merge_parent
 121 # contains 'unknown'.
 122 find_merge_parent()
 123 {
 124   # Find the first line that matches $merge_pattern, do the substitution and
 125   # quit.  Ignore all the other lines.
 126   sed_tmp="s/.*$merge_pattern.*/\\1/"
 127   svn_merge_to=`echo "$2" | sed "$sed_tmp"`
 128   case $svn_merge_to in #(
 129     '' | *[^0-9]*)
 130       fatal "invalid SVN revision '$svn_merge_to' found in $1";;
 131   esac
 132   # Now $svn_merge_to is not necessarily a commit that took part of the
 133   # merge.  For instance, you can merge -r42:51 https://.../branches/foo
 134   # even if the last commit in branch foo is at r46.  So it's utterly
 135   # important that we find the last commit on the branch being merged the
 136   # revision of which must be <= $svn_merge_to (which is 51 in this example).
 137   extract_svn_branch_name "$2"
 138   if test -z "$svn_branch_name"; then
 139     merge_parent='unknown'
 140     return 0
 141   fi
 142   # Create a range to intelligently limit the match of rev-list.  This will
 143   # produce a RE that rules out all the impossible revision numbers (that is,
 144   # the revisions >TO).  e.g:
 145   # 7    -> ([0-7])
 146   # 42   -> (4[0-2]|[0-3][0-9]|[1-9])
 147   # 123  -> (12[0-3]|1[0-1][0-9]|0[0-9][0-9]|[1-9][0-9]{0,1})
 148   # 6951 -> (695[0-1]|69[0-4][0-9]|6[0-8][0-9][0-9]|[0-5][0-9][0-9][0-9]|[1-9][0-9]{0,2})
 149   perl_tmp='$_ = "'"$svn_merge_to"'";
 150             my $l = length($_);
 151             my @r;
 152             foreach my $i (0 .. $l - 1) {
 153               /^(\d*)(\d)(\d{$i})$/;
 154               my ($a, $b, $c) = ($1, int($2), $3);
 155               if ($i != 0) {
 156                 # Avoid pitfalls e.g. 10[0-9] or 0[0-9][0-9] for 101
 157                 next if $b == 0 or ($b == 1 and $a eq "");
 158                 --$b;
 159               }
 160               $b = "[0-$b]" if $b;
 161               $c =~ s/./[0-9]/g;
 162               push(@r, "$a$b$c");
 163             }
 164             push(@r, "[1-9]" . ($l - 2 ? "[0-9]{0," . ($l - 2) . "}" : ""))
 165               if $l > 1;
 166             print "(" . join("|", @r) . ")";'
 167   rev_range=`perl -we "$perl_tmp"`
 168   sed_tmp='s/^ *git-svn-id: .*@\([0-9]*\) [-0-9a-f]*$/\1/p'
 169   svn_merge_parent=`git rev-list --all -1 --header -E \
 170        --grep="^ *git-svn-id: .*/$svn_branch_name@$rev_range [-0-9a-f]*\\$" \
 171        | sed -n "$sed_tmp"`
 172   case $svn_merge_parent in #(
 173     '' | *[^0-9]*) fatal "invalid svn_merge_parent: '$svn_merge_parent'";;
 174   esac
 175   rv=$?
 176   test $rv -eq 0 || fatal "perl returned $rv"
 177   if $opt_verbose; then
 178     if test "$svn_merge_to" -eq "$svn_merge_parent"; then
 179       verb_tmp=
 180     else
 181       verb_tmp=" (in fact r$svn_merge_parent)"
 182     fi
 183   fi
 184   verb "  $1 is merging SVN r$svn_merge_to$verb_tmp from branch $svn_branch_name"
 185   # Now find the sha1 of the merge parent.
 186   merge_parent=`git rev-list --all \
 187                  --grep="^ *git-svn-id: .*@$svn_merge_parent [-0-9a-f]*\\$"`
 188   rv=$?
 189   test $rv -eq 0 || fatal "git rev-list returned $rv"
 190 }
 191
 192 # create_graft <ref> <merge-parent>
 193 # ---------------------------------
 194 # Add <merge-parent> as 2nd parent of the commit designated by <ref>.
 195 create_graft()
 196 {
 197   # --parents will print $1 along with its current parents.
 198   grafted_commit=`git rev-list --no-walk --parents "$1"`
 199   rv=$?
 200   test $rv -eq 0 || fatal "git rev-list returned $rv"
 201   graft_merge_parent=$2
 202
 203   case $grafted_commit in #(
 204     *"$graft_merge_parent"*)
 205       debug "  not grafting commit $1: $graft_merge_parent is already a parent ($grafted_commit)"
 206       return 0;;
 207   esac
 208
 209   graft="$grafted_commit $graft_merge_parent"
 210   existing_graft=`grep "^$1" "$GIT_DIR/info/grafts"`
 211   if test $? -eq 0; then
 212     if test x"$existing_graft" != x"$graft"; then
 213       fatal "$1 is already graft ($existing_graft)\
 214  and the graft is different than what I was going to graft ($graft)"
 215     fi
 216     debug "  not grafting commit $1: already properly grafted"
 217     return 0
 218   fi
 219   debug "  grafting commit $1: add parent $graft_merge_parent"
 220   $opt_dryrun && return 0
 221   nconverted=$((nconverted + 1))
 222   echo >>"$GIT_DIR/info/grafts" "$graft" \
 223   || fatal "Failed to add a graft in $GIT_DIR/info/grafts"
 224 }
 225
 226 # rewrite_history
 227 # ---------------
 228 # Make *all* the grafts part of the actual history.
 229 rewrite_history()
 230 {
 231   $opt_dryrun && return 0
 232   git filter-branch --parent-filter cat -- --all \
 233     || fatal "git filter-branch returned $?"
 234 }
 235
 236 # doit <REF>
 237 # ----------
 238 # Find all the merge mentionned in the commit messages and make them become
 239 # real Git merges.
 240 doit()
 241 {
 242   refspec=$1
 243   verb " >> Processing merges in the history of $refspec"
 244
 245   git rev-list --no-walk "$refspec" >/dev/null \
 246     || fatal "'$refspec' does not seem to be a valid refspec"
 247
 248   git rev-list --grep="$merge_pattern" "$refspec" >"$tmp_buf"
 249   rv=$?
 250   test $rv -eq 0 || fatal "git rev-list failed and returned $rv"
 251   while read commit; do
 252     merge_log=`git log --no-walk "$commit"`
 253     rv=$?
 254     test $rv -eq 0 || fatal "git log returned $rv"
 255     merge_line=`echo "$merge_log" | sed "/$merge_pattern/!d;//q"`
 256
 257     # Maybe skip the commit if it matches $exclude_pattern or
 258     # $log_exclude_pattern (in which case it's not a merge)
 259     if $has_exclude \
 260     && echo "$merge_line" | grep -- "$exclude_pattern" >/dev/null; then
 261       verb "  skipping $commit whose log merge-line is: $merge_line"
 262       continue
 263     fi
 264     if $has_log_exclude \
 265     && echo "$merge_log" | grep -- "$log_exclude_pattern" >/dev/null; then
 266       verb "  skipping $commit whose log is:"
 267       $opt_verbose && echo "$merge_log" | sed 's/^/  | /'
 268       continue
 269     fi
 270
 271     nmerge=$(($nmerge + 1))
 272     verb "  $commit is a merge commit, log says:
 273   | $merge_line"
 274
 275     find_merge_parent "$commit" "$merge_line"
 276
 277     case $merge_parent in #(
 278       unknown)
 279         warn "could not find the merge parent of $commit"
 280         continue;; #(
 281       '' | *[^0-9a-f]*)
 282         fatal "invalid merge_parent: '$merge_parent'";;
 283     esac
 284
 285     create_graft "$commit" "$merge_parent"
 286     test $? -eq 0 || fatal "failed to create a graft for commit $commit"
 287   done <"$tmp_buf"
 288 }
 289
 290 # ------------------ #
 291 # `main' starts here #
 292 # ------------------ #
 293
 294 test -d "$TMPDIR" || fatal "TMPDIR='$TMPDIR' is not a directory"
 295 test -w "$TMPDIR" || fatal "TMPDIR='$TMPDIR' is not writable"
 296 tmp_buf=`mktemp "$TMPDIR/$me.XXXXXX"`
 297 # Clean up temp file upon exit.
 298 trap "exit_status=$?; rm -f $tmp_buf; exit \$exit_status" 0
 299
 300 # Parse the options passed to the script.
 301 # Initialize the defaults
 302 opt_all=false
 303 opt_debug=false
 304 opt_dryrun=false
 305 opt_remote=false
 306 opt_verbose=false
 307
 308 while test $# != 0
 309 do
 310   case $1 in #(
 311     -a | --all)
 312       opt_all=:;; #(
 313     -d | --debug)
 314       opt_debug=:;; #(
 315     -n | --dry-run)
 316       opt_dryrun=:;; #(
 317     -r | --remote)
 318       opt_remote=:;; #(
 319     -v | --verbose)
 320       opt_verbose=:;; #(
 321     --)
 322       shift; break;; #(
 323     -*)
 324       usage;; #(
 325   esac
 326   shift
 327 done
 328
 329 if $opt_all; then
 330   git for-each-ref --shell --format='ref=%(refname)' refs/heads >"$tmp_buf"
 331   rv=$?
 332   test $rv -eq 0 || fatal "git for-each-ref failed and returned $rv"
 333   while read line
 334   do
 335     eval "$line"
 336     set "$@" "$ref"
 337   done <"$tmp_buf"
 338 fi
 339
 340 if $opt_remote; then
 341   git for-each-ref --shell --format='ref=%(refname)' refs/remotes >"$tmp_buf"
 342   rv=$?
 343   test $rv -eq 0 || fatal "git for-each-ref failed and returned $rv"
 344   while read line
 345   do
 346     eval "$line"
 347     branch=`basename "$ref"`
 348     case $branch in #(
 349       HEAD) # Skip branches named `HEAD' (which does happen)
 350         continue;; # because they create ambiguities.
 351     esac
 352     # if the local $branch does not already exist, we create one
 353     exists=`git rev-list --no-walk refs/heads/"$branch" 2>/dev/null`
 354     if test -z "$exists"; then # the $branch does not locally exist
 355       verb "creating branch '$branch' from '$ref'"
 356       git branch "$branch" "$ref" \
 357         || fatal "could not create branch '$branch' from '$ref'"
 358     else # there already is a local $branch
 359       sha1=`git rev-list --no-walk "$ref"`
 360       # Maybe the existing local $branch is identical to the remote $ref?
 361       if test "$sha1" = "$exists"; then # OK, local = remote
 362         verb "branch '$branch' is already properly initialized to '$ref'"
 363         set "$@" "refs/heads/$branch"
 364       else # KO, local != remote
 365         warn "there already exists a local branch '$branch'
 366 and it is at $exists whereas the remote branch '$ref'
 367 is at $sha1 so I'm skipping it..."
 368       fi
 369     fi
 370   done <"$tmp_buf"
 371 fi
 372
 373 # No refspec given => work on HEAD
 374 test -z "$*" && set HEAD
 375
 376 if test -z "$exclude_pattern"; then
 377   has_exclude=false
 378 else
 379   has_exclude=:
 380 fi
 381
 382 if test -z "$log_exclude_pattern"; then
 383   has_log_exclude=false
 384 else
 385   has_log_exclude=:
 386 fi
 387
 388 totalmerge=0
 389 totalconverted=0
 390 for refspec
 391 do
 392   nconverted=0
 393   nmerge=0
 394   doit "$refspec"
 395   echo ">>> processed $nconverted/$nmerge merges in $refspec"
 396   totalmerge=$(($totalmerge + $nmerge))
 397   totalconverted=$(($totalconverted + $nconverted))
 398 done
 399
 400 rewrite_history
 401
 402 echo "Done.  Processed $totalconverted/$totalmerge merges"
 403 test $warnings -eq 0 || warn "job completed with $warnings warnings"