]> git.ipfire.org Git - thirdparty/git.git/blob - git-cvsexportcommit.perl
Merge branch 'gf/maint-sh-setup-nongit-ok' into maint
[thirdparty/git.git] / git-cvsexportcommit.perl
1 #!/usr/bin/perl -w
2
3 use strict;
4 use Getopt::Std;
5 use File::Temp qw(tempdir);
6 use Data::Dumper;
7 use File::Basename qw(basename dirname);
8 use File::Spec;
9 use Git;
10
11 our ($opt_h, $opt_P, $opt_p, $opt_v, $opt_c, $opt_f, $opt_a, $opt_m, $opt_d, $opt_u, $opt_w, $opt_W, $opt_k);
12
13 getopts('uhPpvcfkam:d:w:W');
14
15 $opt_h && usage();
16
17 die "Need at least one commit identifier!" unless @ARGV;
18
19 # Get git-config settings
20 my $repo = Git->repository();
21 $opt_w = $repo->config('cvsexportcommit.cvsdir') unless defined $opt_w;
22
23 if ($opt_w || $opt_W) {
24 # Remember where GIT_DIR is before changing to CVS checkout
25 unless ($ENV{GIT_DIR}) {
26 # No GIT_DIR set. Figure it out for ourselves
27 my $gd =`git-rev-parse --git-dir`;
28 chomp($gd);
29 $ENV{GIT_DIR} = $gd;
30 }
31 # Make sure GIT_DIR is absolute
32 $ENV{GIT_DIR} = File::Spec->rel2abs($ENV{GIT_DIR});
33 }
34
35 if ($opt_w) {
36 if (! -d $opt_w."/CVS" ) {
37 die "$opt_w is not a CVS checkout";
38 }
39 chdir $opt_w or die "Cannot change to CVS checkout at $opt_w";
40 }
41 unless ($ENV{GIT_DIR} && -r $ENV{GIT_DIR}){
42 die "GIT_DIR is not defined or is unreadable";
43 }
44
45
46 my @cvs;
47 if ($opt_d) {
48 @cvs = ('cvs', '-d', $opt_d);
49 } else {
50 @cvs = ('cvs');
51 }
52
53 # resolve target commit
54 my $commit;
55 $commit = pop @ARGV;
56 $commit = safe_pipe_capture('git-rev-parse', '--verify', "$commit^0");
57 chomp $commit;
58 if ($?) {
59 die "The commit reference $commit did not resolve!";
60 }
61
62 # resolve what parent we want
63 my $parent;
64 if (@ARGV) {
65 $parent = pop @ARGV;
66 $parent = safe_pipe_capture('git-rev-parse', '--verify', "$parent^0");
67 chomp $parent;
68 if ($?) {
69 die "The parent reference did not resolve!";
70 }
71 }
72
73 # find parents from the commit itself
74 my @commit = safe_pipe_capture('git-cat-file', 'commit', $commit);
75 my @parents;
76 my $committer;
77 my $author;
78 my $stage = 'headers'; # headers, msg
79 my $title;
80 my $msg = '';
81
82 foreach my $line (@commit) {
83 chomp $line;
84 if ($stage eq 'headers' && $line eq '') {
85 $stage = 'msg';
86 next;
87 }
88
89 if ($stage eq 'headers') {
90 if ($line =~ m/^parent (\w{40})$/) { # found a parent
91 push @parents, $1;
92 } elsif ($line =~ m/^author (.+) \d+ [-+]\d+$/) {
93 $author = $1;
94 } elsif ($line =~ m/^committer (.+) \d+ [-+]\d+$/) {
95 $committer = $1;
96 }
97 } else {
98 $msg .= $line . "\n";
99 unless ($title) {
100 $title = $line;
101 }
102 }
103 }
104
105 my $noparent = "0000000000000000000000000000000000000000";
106 if ($parent) {
107 my $found;
108 # double check that it's a valid parent
109 foreach my $p (@parents) {
110 if ($p eq $parent) {
111 $found = 1;
112 last;
113 }; # found it
114 }
115 die "Did not find $parent in the parents for this commit!" if !$found and !$opt_P;
116 } else { # we don't have a parent from the cmdline...
117 if (@parents == 1) { # it's safe to get it from the commit
118 $parent = $parents[0];
119 } elsif (@parents == 0) { # there is no parent
120 $parent = $noparent;
121 } else { # cannot choose automatically from multiple parents
122 die "This commit has more than one parent -- please name the parent you want to use explicitly";
123 }
124 }
125
126 my $go_back_to = 0;
127
128 if ($opt_W) {
129 $opt_v && print "Resetting to $parent\n";
130 $go_back_to = `git symbolic-ref HEAD 2> /dev/null ||
131 git rev-parse HEAD` || die "Could not determine current branch";
132 system("git checkout -q $parent^0") && die "Could not check out $parent^0";
133 }
134
135 $opt_v && print "Applying to CVS commit $commit from parent $parent\n";
136
137 # grab the commit message
138 open(MSG, ">.msg") or die "Cannot open .msg for writing";
139 if ($opt_m) {
140 print MSG $opt_m;
141 }
142 print MSG $msg;
143 if ($opt_a) {
144 print MSG "\n\nAuthor: $author\n";
145 if ($author ne $committer) {
146 print MSG "Committer: $committer\n";
147 }
148 }
149 close MSG;
150
151 if ($parent eq $noparent) {
152 `git-diff-tree --binary -p --root $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
153 } else {
154 `git-diff-tree --binary -p $parent $commit >.cvsexportcommit.diff`;# || die "Cannot diff";
155 }
156
157 ## apply non-binary changes
158
159 # In pedantic mode require all lines of context to match. In normal
160 # mode, be compatible with diff/patch: assume 3 lines of context and
161 # require at least one line match, i.e. ignore at most 2 lines of
162 # context, like diff/patch do by default.
163 my $context = $opt_p ? '' : '-C1';
164
165 print "Checking if patch will apply\n";
166
167 my @stat;
168 open APPLY, "GIT_DIR= git-apply $context --summary --numstat<.cvsexportcommit.diff|" || die "cannot patch";
169 @stat=<APPLY>;
170 close APPLY || die "Cannot patch";
171 my (@bfiles,@files,@afiles,@dfiles);
172 chomp @stat;
173 foreach (@stat) {
174 push (@bfiles,$1) if m/^-\t-\t(.*)$/;
175 push (@files, $1) if m/^-\t-\t(.*)$/;
176 push (@files, $1) if m/^\d+\t\d+\t(.*)$/;
177 push (@afiles,$1) if m/^ create mode [0-7]+ (.*)$/;
178 push (@dfiles,$1) if m/^ delete mode [0-7]+ (.*)$/;
179 }
180 map { s/^"(.*)"$/$1/g } @bfiles,@files;
181 map { s/\\([0-7]{3})/sprintf('%c',oct $1)/eg } @bfiles,@files;
182
183 # check that the files are clean and up to date according to cvs
184 my $dirty;
185 my @dirs;
186 foreach my $p (@afiles) {
187 my $path = dirname $p;
188 while (!-d $path and ! grep { $_ eq $path } @dirs) {
189 unshift @dirs, $path;
190 $path = dirname $path;
191 }
192 }
193
194 # ... check dirs,
195 foreach my $d (@dirs) {
196 if (-e $d) {
197 $dirty = 1;
198 warn "$d exists and is not a directory!\n";
199 }
200 }
201
202 # ... query status of all files that we have a directory for and parse output of 'cvs status' to %cvsstat.
203 my @canstatusfiles;
204 foreach my $f (@files) {
205 my $path = dirname $f;
206 next if (grep { $_ eq $path } @dirs);
207 push @canstatusfiles, $f;
208 }
209
210 my %cvsstat;
211 if (@canstatusfiles) {
212 if ($opt_u) {
213 my @updated = xargs_safe_pipe_capture([@cvs, 'update'], @canstatusfiles);
214 print @updated;
215 }
216 # "cvs status" reorders the parameters, notably when there are multiple
217 # arguments with the same basename. So be precise here.
218
219 my %added = map { $_ => 1 } @afiles;
220 my %todo = map { $_ => 1 } @canstatusfiles;
221
222 while (%todo) {
223 my @canstatusfiles2 = ();
224 my %fullname = ();
225 foreach my $name (keys %todo) {
226 my $basename = basename($name);
227
228 # CVS reports files that don't exist in the current revision as
229 # "no file $basename" in its "status" output, so we should
230 # anticipate that. Totally unknown files will have a status
231 # "Unknown". However, if they exist in the Attic, their status
232 # will be "Up-to-date" (this means they were added once but have
233 # been removed).
234 $basename = "no file $basename" if $added{$basename};
235
236 $basename =~ s/^\s+//;
237 $basename =~ s/\s+$//;
238
239 if (!exists($fullname{$basename})) {
240 $fullname{$basename} = $name;
241 push (@canstatusfiles2, $name);
242 delete($todo{$name});
243 }
244 }
245 my @cvsoutput;
246 @cvsoutput = xargs_safe_pipe_capture([@cvs, 'status'], @canstatusfiles2);
247 foreach my $l (@cvsoutput) {
248 chomp $l;
249 next unless
250 my ($file, $status) = $l =~ /^File:\s+(.*\S)\s+Status: (.*)$/;
251
252 my $fullname = $fullname{$file};
253 print STDERR "Huh? Status '$status' reported for unexpected file '$file'\n"
254 unless defined $fullname;
255
256 # This response means the file does not exist except in
257 # CVS's attic, so set the status accordingly
258 $status = "In-attic"
259 if $file =~ /^no file /
260 && $status eq 'Up-to-date';
261
262 $cvsstat{$fullname{$file}} = $status
263 if defined $fullname{$file};
264 }
265 }
266 }
267
268 # ... Validate that new files have the correct status
269 foreach my $f (@afiles) {
270 next unless defined(my $stat = $cvsstat{$f});
271
272 # This means the file has never been seen before
273 next if $stat eq 'Unknown';
274
275 # This means the file has been seen before but was removed
276 next if $stat eq 'In-attic';
277
278 $dirty = 1;
279 warn "File $f is already known in your CVS checkout -- perhaps it has been added by another user. Or this may indicate that it exists on a different branch. If this is the case, use -f to force the merge.\n";
280 warn "Status was: $cvsstat{$f}\n";
281 }
282
283 # ... validate known files.
284 foreach my $f (@files) {
285 next if grep { $_ eq $f } @afiles;
286 # TODO:we need to handle removed in cvs
287 unless (defined ($cvsstat{$f}) and $cvsstat{$f} eq "Up-to-date") {
288 $dirty = 1;
289 warn "File $f not up to date but has status '$cvsstat{$f}' in your CVS checkout!\n";
290 }
291
292 # Depending on how your GIT tree got imported from CVS you may
293 # have a conflict between expanded keywords in your CVS tree and
294 # unexpanded keywords in the patch about to be applied.
295 if ($opt_k) {
296 my $orig_file ="$f.orig";
297 rename $f, $orig_file;
298 open(FILTER_IN, "<$orig_file") or die "Cannot open $orig_file\n";
299 open(FILTER_OUT, ">$f") or die "Cannot open $f\n";
300 while (<FILTER_IN>)
301 {
302 my $line = $_;
303 $line =~ s/\$([A-Z][a-z]+):[^\$]+\$/\$$1\$/g;
304 print FILTER_OUT $line;
305 }
306 close FILTER_IN;
307 close FILTER_OUT;
308 }
309 }
310
311 if ($dirty) {
312 if ($opt_f) { warn "The tree is not clean -- forced merge\n";
313 $dirty = 0;
314 } else {
315 die "Exiting: your CVS tree is not clean for this merge.";
316 }
317 }
318
319 print "Applying\n";
320 if ($opt_W) {
321 system("git checkout -q $commit^0") && die "cannot patch";
322 } else {
323 `GIT_DIR= git-apply $context --summary --numstat --apply <.cvsexportcommit.diff` || die "cannot patch";
324 }
325
326 print "Patch applied successfully. Adding new files and directories to CVS\n";
327 my $dirtypatch = 0;
328
329 #
330 # We have to add the directories in order otherwise we will have
331 # problems when we try and add the sub-directory of a directory we
332 # have not added yet.
333 #
334 # Luckily this is easy to deal with by sorting the directories and
335 # dealing with the shortest ones first.
336 #
337 @dirs = sort { length $a <=> length $b} @dirs;
338
339 foreach my $d (@dirs) {
340 if (system(@cvs,'add',$d)) {
341 $dirtypatch = 1;
342 warn "Failed to cvs add directory $d -- you may need to do it manually";
343 }
344 }
345
346 foreach my $f (@afiles) {
347 if (grep { $_ eq $f } @bfiles) {
348 system(@cvs, 'add','-kb',$f);
349 } else {
350 system(@cvs, 'add', $f);
351 }
352 if ($?) {
353 $dirtypatch = 1;
354 warn "Failed to cvs add $f -- you may need to do it manually";
355 }
356 }
357
358 foreach my $f (@dfiles) {
359 system(@cvs, 'rm', '-f', $f);
360 if ($?) {
361 $dirtypatch = 1;
362 warn "Failed to cvs rm -f $f -- you may need to do it manually";
363 }
364 }
365
366 print "Commit to CVS\n";
367 print "Patch title (first comment line): $title\n";
368 my @commitfiles = map { unless (m/\s/) { '\''.$_.'\''; } else { $_; }; } (@files);
369 my $cmd = join(' ', @cvs)." commit -F .msg @commitfiles";
370
371 if ($dirtypatch) {
372 print "NOTE: One or more hunks failed to apply cleanly.\n";
373 print "You'll need to apply the patch in .cvsexportcommit.diff manually\n";
374 print "using a patch program. After applying the patch and resolving the\n";
375 print "problems you may commit using:";
376 print "\n cd \"$opt_w\"" if $opt_w;
377 print "\n $cmd\n";
378 print "\n git checkout $go_back_to\n" if $go_back_to;
379 print "\n";
380 exit(1);
381 }
382
383 if ($opt_c) {
384 print "Autocommit\n $cmd\n";
385 print xargs_safe_pipe_capture([@cvs, 'commit', '-F', '.msg'], @files);
386 if ($?) {
387 die "Exiting: The commit did not succeed";
388 }
389 print "Committed successfully to CVS\n";
390 # clean up
391 unlink(".msg");
392 } else {
393 print "Ready for you to commit, just run:\n\n $cmd\n";
394 }
395
396 # clean up
397 unlink(".cvsexportcommit.diff");
398
399 if ($opt_W) {
400 system("git checkout $go_back_to") && die "cannot move back to $go_back_to";
401 if (!($go_back_to =~ /^[0-9a-fA-F]{40}$/)) {
402 system("git symbolic-ref HEAD $go_back_to") &&
403 die "cannot move back to $go_back_to";
404 }
405 }
406
407 # CVS version 1.11.x and 1.12.x sleeps the wrong way to ensure the timestamp
408 # used by CVS and the one set by subsequence file modifications are different.
409 # If they are not different CVS will not detect changes.
410 sleep(1);
411
412 sub usage {
413 print STDERR <<END;
414 Usage: GIT_DIR=/path/to/.git git cvsexportcommit [-h] [-p] [-v] [-c] [-f] [-u] [-k] [-w cvsworkdir] [-m msgprefix] [ parent ] commit
415 END
416 exit(1);
417 }
418
419 # An alternative to `command` that allows input to be passed as an array
420 # to work around shell problems with weird characters in arguments
421 # if the exec returns non-zero we die
422 sub safe_pipe_capture {
423 my @output;
424 if (my $pid = open my $child, '-|') {
425 @output = (<$child>);
426 close $child or die join(' ',@_).": $! $?";
427 } else {
428 exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
429 }
430 return wantarray ? @output : join('',@output);
431 }
432
433 sub xargs_safe_pipe_capture {
434 my $MAX_ARG_LENGTH = 65536;
435 my $cmd = shift;
436 my @output;
437 my $output;
438 while(@_) {
439 my @args;
440 my $length = 0;
441 while(@_ && $length < $MAX_ARG_LENGTH) {
442 push @args, shift;
443 $length += length($args[$#args]);
444 }
445 if (wantarray) {
446 push @output, safe_pipe_capture(@$cmd, @args);
447 }
448 else {
449 $output .= safe_pipe_capture(@$cmd, @args);
450 }
451 }
452 return wantarray ? @output : $output;
453 }