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   svn_merge_parent=`git rev-list --all --header \
 143              --grep="^ *git-svn-id: .*/$svn_branch_name@[0-9]* [-0-9a-f]*\\$" |
 144              perl -we 'my $rev = 0; my $t = 0;
 145              while(<>) {
 146                m{^ *git-svn-id: .*/\Q'"$svn_branch_name"'\E@([0-9]+) [-0-9a-f]+$}
 147                  or next;
 148                $t = int($1);
 149                $rev = $t if $rev < $t and $t <= '"$svn_merge_to"';
 150              }
 151              print $rev'`
 152   rv=$?
 153   test $rv -eq 0 || fatal "perl returned $rv"
 154   if $opt_verbose; then
 155     if test "$svn_merge_to" -eq "$svn_merge_parent"; then
 156       verb_tmp=
 157     else
 158       verb_tmp=" (in fact r$svn_merge_parent)"
 159     fi
 160   fi
 161   verb "  $1 is merging SVN r$svn_merge_to$verb_tmp from branch $svn_branch_name"
 162   # Now find the sha1 of the merge parent.
 163   merge_parent=`git rev-list --all \
 164                  --grep="^ *git-svn-id: .*@$svn_merge_parent [-0-9a-f]*\\$"`
 165   rv=$?
 166   test $rv -eq 0 || fatal "git rev-list returned $rv"
 167 }
 168
 169 # create_graft <ref> <merge-parent>
 170 # ---------------------------------
 171 # Add <merge-parent> as 2nd parent of the commit designated by <ref>.
 172 create_graft()
 173 {
 174   # --parents will print $1 along with its current parents.
 175   grafted_commit=`git rev-list --no-walk --parents "$1"`
 176   rv=$?
 177   test $rv -eq 0 || fatal "git rev-list returned $rv"
 178   graft_merge_parent=$2
 179
 180   case $grafted_commit in #(
 181     *"$graft_merge_parent"*)
 182       debug "  not grafting commit $1: $graft_merge_parent is already a parent ($grafted_commit)"
 183       return 0;;
 184   esac
 185
 186   graft="$grafted_commit $graft_merge_parent"
 187   existing_graft=`grep "^$1" "$GIT_DIR/info/grafts"`
 188   if test $? -eq 0; then
 189     if test x"$existing_graft" != x"$graft"; then
 190       fatal "$1 is already graft ($existing_graft)\
 191  and the graft is different than what I was going to graft ($graft)"
 192     fi
 193     debug "  not grafting commit $1: already properly grafted"
 194     return 0
 195   fi
 196   debug "  grafting commit $1: add parent $graft_merge_parent"
 197   $opt_dryrun && return 0
 198   nconverted=$((nconverted + 1))
 199   echo >>"$GIT_DIR/info/grafts" "$graft" \
 200   || fatal "Failed to add a graft in $GIT_DIR/info/grafts"
 201 }
 202
 203 # rewrite_history
 204 # ---------------
 205 # Make *all* the grafts part of the actual history.
 206 rewrite_history()
 207 {
 208   $opt_dryrun && return 0
 209   git filter-branch --parent-filter cat -- --all
 210 }
 211
 212 # doit <REF>
 213 # ----------
 214 # Find all the merge mentionned in the commit messages and make them become
 215 # real Git merges.
 216 doit()
 217 {
 218   refspec=$1
 219   verb " >> Processing merges in the history of $refspec"
 220
 221   git rev-list --no-walk "$refspec" >/dev/null \
 222     || fatal "'$refspec' does not seem to be a valid refspec"
 223
 224   git rev-list --grep="$merge_pattern" "$refspec" >"$tmp_buf"
 225   rv=$?
 226   test $rv -eq 0 || fatal "git rev-list failed and returned $rv"
 227   while read commit; do
 228     merge_log=`git log --no-walk "$commit"`
 229     rv=$?
 230     test $rv -eq 0 || fatal "git log returned $rv"
 231     merge_line=`echo "$merge_log" | sed "/$merge_pattern/!d;//q"`
 232
 233     # Maybe skip the commit if it matches $exclude_pattern or
 234     # $log_exclude_pattern (in which case it's not a merge)
 235     if $has_exclude \
 236     && echo "$merge_line" | grep -- "$exclude_pattern" >/dev/null; then
 237       verb "  skipping $commit whose log merge-line is: $merge_line"
 238       continue
 239     fi
 240     if $has_log_exclude \
 241     && echo "$merge_log" | grep -- "$log_exclude_pattern" >/dev/null; then
 242       verb "  skipping $commit whose log is:"
 243       $opt_verbose && echo "$merge_log" | sed 's/^/  | /'
 244       continue
 245     fi
 246
 247     nmerge=$(($nmerge + 1))
 248     verb "  $commit is a merge commit, log says:
 249   | $merge_line"
 250
 251     find_merge_parent "$commit" "$merge_line"
 252
 253     case $merge_parent in #(
 254       unknown)
 255         warn "could not find the merge parent of $commit"
 256         continue;; #(
 257       '' | *[^0-9a-f]*)
 258         fatal "invalid merge_parent: '$merge_parent'";;
 259     esac
 260
 261     create_graft "$commit" "$merge_parent"
 262     test $? -eq 0 || fatal "failed to create a graft for commit $commit"
 263   done <"$tmp_buf"
 264 }
 265
 266 # ------------------ #
 267 # `main' starts here #
 268 # ------------------ #
 269
 270 test -d "$TMPDIR" || fatal "TMPDIR='$TMPDIR' is not a directory"
 271 test -w "$TMPDIR" || fatal "TMPDIR='$TMPDIR' is not writable"
 272 tmp_buf=`mktemp "$TMPDIR/$me.XXXXXX"`
 273 # Clean up temp file upon exit.
 274 trap "exit_status=$?; rm -f $tmp_buf; exit \$exit_status" 0
 275
 276 # Parse the options passed to the script.
 277 # Initialize the defaults
 278 opt_all=false
 279 opt_debug=false
 280 opt_dryrun=false
 281 opt_remote=false
 282 opt_verbose=false
 283
 284 while test $# != 0
 285 do
 286   case $1 in #(
 287     -a | --all)
 288       opt_all=:;; #(
 289     -d | --debug)
 290       opt_debug=:;; #(
 291     -n | --dry-run)
 292       opt_dryrun=:;; #(
 293     -r | --remote)
 294       opt_remote=:;; #(
 295     -v | --verbose)
 296       opt_verbose=:;; #(
 297     --)
 298       shift; break;; #(
 299     -*)
 300       usage;; #(
 301   esac
 302   shift
 303 done
 304
 305 if $opt_all; then
 306   git for-each-ref --shell --format='ref=%(refname)' refs/heads >"$tmp_buf"
 307   rv=$?
 308   test $rv -eq 0 || fatal "git for-each-ref failed and returned $rv"
 309   while read line
 310   do
 311     eval "$line"
 312     set "$@" "$ref"
 313   done <"$tmp_buf"
 314 fi
 315
 316 if $opt_remote; then
 317   git for-each-ref --shell --format='ref=%(refname)' refs/remotes >"$tmp_buf"
 318   rv=$?
 319   test $rv -eq 0 || fatal "git for-each-ref failed and returned $rv"
 320   while read line
 321   do
 322     eval "$line"
 323     branch=`basename "$ref"`
 324     case $branch in #(
 325       HEAD) # Skip branches named `HEAD' (which does happen)
 326         continue;; # because they create ambiguities.
 327     esac
 328     # if the local $branch does not already exist, we create one
 329     exists=`git rev-list --no-walk refs/heads/"$branch" 2>/dev/null`
 330     if test -z "$exists"; then # the $branch does not locally exist
 331       verb "creating branch '$branch' from '$ref'"
 332       git branch "$branch" "$ref" \
 333         || fatal "could not create branch '$branch' from '$ref'"
 334     else # there already is a local $branch
 335       sha1=`git rev-list --no-walk "$ref"`
 336       # Maybe the existing local $branch is identical to the remote $ref?
 337       if test "$sha1" = "$exists"; then # OK, local = remote
 338         verb "branch '$branch' is already properly initialized to '$ref'"
 339         set "$@" "refs/heads/$branch"
 340       else # KO, local != remote
 341         warn "there already exists a local branch '$branch'
 342 and it is at $exists whereas the remote branch '$ref'
 343 is at $sha1 so I'm skipping it..."
 344       fi
 345     fi
 346   done <"$tmp_buf"
 347 fi
 348
 349 # No refspec given => work on HEAD
 350 test -z "$*" && set HEAD
 351
 352 if test -z "$exclude_pattern"; then
 353   has_exclude=false
 354 else
 355   has_exclude=:
 356 fi
 357
 358 if test -z "$log_exclude_pattern"; then
 359   has_log_exclude=false
 360 else
 361   has_log_exclude=:
 362 fi
 363
 364 for refspec
 365 do
 366   nconverted=0
 367   nmerge=0
 368   doit "$refspec"
 369   echo ">>> processed $nconverted/$nmerge merges in $refspec"
 370 done
 371
 372 rewrite_history
 373
 374 test $warnings -eq 0 || warn "job completed with $warnings warnings"