contrib/fast-import/import-directories.perl

   1 #!/usr/bin/perl -w
   2 #
   3 # Copyright 2008-2009 Peter Krefting <peter@softwolves.pp.se>
   4 #
   5 # ------------------------------------------------------------------------
   6 #
   7 # This program is free software; you can redistribute it and/or modify
   8 # it under the terms of the GNU General Public License as published by
   9 # the Free Software Foundation.
  10 #
  11 # This program is distributed in the hope that it will be useful,
  12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 # GNU General Public License for more details.
  15 #
  16 # You should have received a copy of the GNU General Public License
  17 # along with this program; if not, write to the Free Software
  18 # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  19 #
  20 # ------------------------------------------------------------------------
  21
  22 =pod
  23
  24 =head1 NAME
  25
  26 import-directories - Import bits and pieces to Git.
  27
  28 =head1 SYNOPSIS
  29
  30 B<import-directories.perl> F<configfile> F<outputfile>
  31
  32 =head1 DESCRIPTION
  33
  34 Script to import arbitrary projects version controlled by the "copy the
  35 source directory to a new location and edit it there"-version controlled
  36 projects into version control. Handles projects with arbitrary branching
  37 and version trees, taking a file describing the inputs and generating a
  38 file compatible with the L<git-fast-import(1)> format.
  39
  40 =head1 CONFIGURATION FILE
  41
  42 =head2 Format
  43
  44 The configuration file is based on the standard I<.ini> format.
  45
  46  ; Comments start with semi-colons
  47  [section]
  48  key=value
  49
  50 Please see below for information on how to escape special characters.
  51
  52 =head2 Global configuration
  53
  54 Global configuration is done in the B<[config]> section, which should be
  55 the first section in the file. Configuration can be changed by
  56 repeating configuration sections later on.
  57
  58  [config]
  59  ; configure conversion of CRLFs. "convert" means that all CRLFs
  60  ; should be converted into LFs (suitable for the core.autocrlf
  61  ; setting set to true in Git). "none" means that all data is
  62  ; treated as binary.
  63  crlf=convert
  64
  65 =head2 Revision configuration
  66
  67 Each revision that is to be imported is described in three
  68 sections. Revisions should be defined in topological order, so
  69 that a revision's parent has always been defined when a new revision
  70 is introduced. All the sections for one revision must be defined
  71 before defining the next revision.
  72
  73 Each revision is assigned a unique numerical identifier. The
  74 numbers do not need to be consecutive, nor monotonically
  75 increasing.
  76
  77 For instance, if your configuration file contains only the two
  78 revisions 4711 and 42, where 4711 is the initial commit, the
  79 only requirement is that 4711 is completely defined before 42.
  80
  81 =pod
  82
  83 =head3 Revision description section
  84
  85 A section whose section name is just an integer gives meta-data
  86 about the revision.
  87
  88  [3]
  89  ; author sets the author of the revisions
  90  author=Peter Krefting <peter@softwolves.pp.se>
  91  ; branch sets the branch that the revision should be committed to
  92  branch=master
  93  ; parent describes the revision that is the parent of this commit
  94  ; (optional)
  95  parent=1
  96  ; merges describes a revision that is merged into this commit
  97  ; (optional; can be repeated)
  98  merges=2
  99  ; selects one file to take the timestamp from
 100  ; (optional; if unspecified, the most recent file from the .files
 101  ;  section is used)
 102  timestamp=3/source.c
 103
 104 =head3 Revision contents section
 105
 106 A section whose section name is an integer followed by B<.files>
 107 describe all the files included in this revision. If a file that
 108 was available previously is not included in this revision, it will
 109 be removed.
 110
 111 If an on-disk revision is incomplete, you can point to files from
 112 a previous revision. There are no restriction as to where the source
 113 files are located, nor to the names of them.
 114
 115  [3.files]
 116  ; the key is the path inside the repository, the value is the path
 117  ; as seen from the importer script.
 118  source.c=ver-3.00/source.c
 119  source.h=ver-2.99/source.h
 120  readme.txt=ver-3.00/introduction to the project.txt
 121
 122 File names are treated as byte strings (but please see below on
 123 quoting rules), and should be stored in the configuration file in
 124 the encoding that should be used in the generated repository.
 125
 126 =head3 Revision commit message section
 127
 128 A section whose section name is an integer followed by B<.message>
 129 gives the commit message. This section is read verbatim, up until
 130 the beginning of the next section. As such, a commit message may not
 131 contain a line that begins with an opening square bracket ("[") and
 132 ends with a closing square bracket ("]"), unless they are surrounded
 133 by whitespace or other characters.
 134
 135  [3.message]
 136  Implement foobar.
 137  ; trailing blank lines are ignored.
 138
 139 =cut
 140
 141 # Globals
 142 use strict;
 143 use integer;
 144 my $crlfmode = 0;
 145 my @revs;
 146 my (%revmap, %message, %files, %author, %branch, %parent, %merges, %time, %timesource);
 147 my $sectiontype = 0;
 148 my $rev = 0;
 149 my $mark = 1;
 150
 151 # Check command line
 152 if ($#ARGV < 1 || $ARGV[0] =~ /^--?h/)
 153 {
 154     exec('perldoc', $0);
 155     exit 1;
 156 }
 157
 158 # Open configuration
 159 my $config = $ARGV[0];
 160 open CFG, '<', $config or die "Cannot open configuration file \"$config\": ";
 161
 162 # Open output
 163 my $output = $ARGV[1];
 164 open OUT, '>', $output or die "Cannot create output file \"$output\": ";
 165 binmode OUT;
 166
 167 LINE: while (my $line = <CFG>)
 168 {
 169         $line =~ s/\r?\n$//;
 170         next LINE if $sectiontype != 4 && $line eq '';
 171         next LINE if $line =~ /^;/;
 172         my $oldsectiontype = $sectiontype;
 173         my $oldrev = $rev;
 174
 175         # Sections
 176         if ($line =~ m"^\[(config|(\d+)(|\.files|\.message))\]$")
 177         {
 178                 if ($1 eq 'config')
 179                 {
 180                         $sectiontype = 1;
 181                 }
 182                 elsif ($3 eq '')
 183                 {
 184                         $sectiontype = 2;
 185                         $rev = $2;
 186                         # Create a new revision
 187                         die "Duplicate rev: $line\n " if defined $revmap{$rev};
 188                         print "Reading revision $rev\n";
 189                         push @revs, $rev;
 190                         $revmap{$rev} = $mark ++;
 191                         $time{$revmap{$rev}} = 0;
 192                 }
 193                 elsif ($3 eq '.files')
 194                 {
 195                         $sectiontype = 3;
 196                         $rev = $2;
 197                         die "Revision mismatch: $line\n " unless $rev == $oldrev;
 198                 }
 199                 elsif ($3 eq '.message')
 200                 {
 201                         $sectiontype = 4;
 202                         $rev = $2;
 203                         die "Revision mismatch: $line\n " unless $rev == $oldrev;
 204                 }
 205                 else
 206                 {
 207                         die "Internal parse error: $line\n ";
 208                 }
 209                 next LINE;
 210         }
 211
 212         # Parse data
 213         if ($sectiontype != 4)
 214         {
 215                 # Key and value
 216                 if ($line =~ m"^\s*([^\s].*=.*[^\s])\s*$")
 217                 {
 218                         my ($key, $value) = &parsekeyvaluepair($1);
 219                         # Global configuration
 220                         if (1 == $sectiontype)
 221                         {
 222                                 if ($key eq 'crlf')
 223                                 {
 224                                         $crlfmode = 1, next LINE if $value eq 'convert';
 225                                         $crlfmode = 0, next LINE if $value eq 'none';
 226                                 }
 227                                 die "Unknown configuration option: $line\n ";
 228                         }
 229                         # Revision specification
 230                         if (2 == $sectiontype)
 231                         {
 232                                 my $current = $revmap{$rev};
 233                                 $author{$current} = $value, next LINE if $key eq 'author';
 234                                 $branch{$current} = $value, next LINE if $key eq 'branch';
 235                                 $parent{$current} = $value, next LINE if $key eq 'parent';
 236                                 $timesource{$current} = $value, next LINE if $key eq 'timestamp';
 237                                 push(@{$merges{$current}}, $value), next LINE if $key eq 'merges';
 238                                 die "Unknown revision option: $line\n ";
 239                         }
 240                         # Filespecs
 241                         if (3 == $sectiontype)
 242                         {
 243                                 # Add the file and create a marker
 244                                 die "File not found: $line\n " unless -f $value;
 245                                 my $current = $revmap{$rev};
 246                                 ${$files{$current}}{$key} = $mark;
 247                                 my $time = &fileblob($value, $crlfmode, $mark ++);
 248
 249                                 # Update revision timestamp if more recent than other
 250                                 # files seen, or if this is the file we have selected
 251                                 # to take the time stamp from using the "timestamp"
 252                                 # directive.
 253                                 if ((defined $timesource{$current} && $timesource{$current} eq $value)
 254                                     || $time > $time{$current})
 255                                 {
 256                                         $time{$current} = $time;
 257                                 }
 258                         }
 259                 }
 260                 else
 261                 {
 262                         die "Parse error: $line\n ";
 263                 }
 264         }
 265         else
 266         {
 267                 # Commit message
 268                 my $current = $revmap{$rev};
 269                 if (defined $message{$current})
 270                 {
 271                         $message{$current} .= "\n";
 272                 }
 273                 $message{$current} .= $line;
 274         }
 275 }
 276 close CFG;
 277
 278 # Start spewing out data for git-fast-import
 279 foreach my $commit (@revs)
 280 {
 281         # Progress
 282         print OUT "progress Creating revision $commit\n";
 283
 284         # Create commit header
 285         my $mark = $revmap{$commit};
 286
 287         # Branch and commit id
 288         print OUT "commit refs/heads/", $branch{$mark}, "\nmark :", $mark, "\n";
 289
 290         # Author and timestamp
 291         die "No timestamp defined for $commit (no files?)\n" unless defined $time{$mark};
 292         print OUT "committer ", $author{$mark}, " ", $time{$mark}, " +0100\n";
 293
 294         # Commit message
 295         die "No message defined for $commit\n" unless defined $message{$mark};
 296         my $message = $message{$mark};
 297         $message =~ s/\n$//; # Kill trailing empty line
 298         print OUT "data ", length($message), "\n", $message, "\n";
 299
 300         # Parent and any merges
 301         print OUT "from :", $revmap{$parent{$mark}}, "\n" if defined $parent{$mark};
 302         if (defined $merges{$mark})
 303         {
 304                 foreach my $merge (@{$merges{$mark}})
 305                 {
 306                         print OUT "merge :", $revmap{$merge}, "\n";
 307                 }
 308         }
 309
 310         # Output file marks
 311         print OUT "deleteall\n"; # start from scratch
 312         foreach my $file (sort keys %{$files{$mark}})
 313         {
 314                 print OUT "M 644 :", ${$files{$mark}}{$file}, " $file\n";
 315         }
 316         print OUT "\n";
 317 }
 318
 319 # Create one file blob
 320 sub fileblob
 321 {
 322         my ($filename, $crlfmode, $mark) = @_;
 323
 324         # Import the file
 325         print OUT "progress Importing $filename\nblob\nmark :$mark\n";
 326         open FILE, '<', $filename or die "Cannot read $filename\n ";
 327         binmode FILE;
 328         my ($size, $mtime) = (stat(FILE))[7,9];
 329         my $file;
 330         read FILE, $file, $size;
 331         close FILE;
 332         $file =~ s/\r\n/\n/g if $crlfmode;
 333         print OUT "data ", length($file), "\n", $file, "\n";
 334
 335         return $mtime;
 336 }
 337
 338 # Parse a key=value pair
 339 sub parsekeyvaluepair
 340 {
 341 =pod
 342
 343 =head2 Escaping special characters
 344
 345 Key and value strings may be enclosed in quotes, in which case
 346 whitespace inside the quotes is preserved. Additionally, an equal
 347 sign may be included in the key by preceding it with a backslash.
 348 For example:
 349
 350  "key1 "=value1
 351  key2=" value2"
 352  key\=3=value3
 353  key4=value=4
 354  "key5""=value5
 355
 356 Here the first key is "key1 " (note the trailing white-space) and the
 357 second value is " value2" (note the leading white-space). The third
 358 key contains an equal sign "key=3" and so does the fourth value, which
 359 does not need to be escaped. The fifth key contains a trailing quote,
 360 which does not need to be escaped since it is inside a surrounding
 361 quote.
 362
 363 =cut
 364         my $pair = shift;
 365
 366         # Separate key and value by the first non-quoted equal sign
 367         my ($key, $value);
 368         if ($pair =~ /^(.*[^\\])=(.*)$/)
 369         {
 370                 ($key, $value) = ($1, $2)
 371         }
 372         else
 373         {
 374                 die "Parse error: $pair\n ";
 375         }
 376
 377         # Unquote and unescape the key and value separately
 378         return (&unescape($key), &unescape($value));
 379 }
 380
 381 # Unquote and unescape
 382 sub unescape
 383 {
 384         my $string = shift;
 385
 386         # First remove enclosing quotes. Backslash before the trailing
 387         # quote leaves both.
 388         if ($string =~ /^"(.*[^\\])"$/)
 389         {
 390                 $string = $1;
 391         }
 392
 393         # Second remove any backslashes inside the unquoted string.
 394         # For later: Handle special sequences like \t ?
 395         $string =~ s/\\(.)/$1/g;
 396
 397         return $string;
 398 }
 399
 400 __END__
 401
 402 =pod
 403
 404 =head1 EXAMPLES
 405
 406 B<import-directories.perl> F<project.import>
 407
 408 =head1 AUTHOR
 409
 410 Copyright 2008-2009 Peter Krefting E<lt>peter@softwolves.pp.se>
 411
 412 This program is free software; you can redistribute it and/or modify
 413 it under the terms of the GNU General Public License as published by
 414 the Free Software Foundation.
 415
 416 =cut