git-cvsserver.perl

   1 #!/usr/bin/perl
   2
   3 ####
   4 #### This application is a CVS emulation layer for git.
   5 #### It is intended for clients to connect over SSH.
   6 #### See the documentation for more details.
   7 ####
   8 #### Copyright The Open University UK - 2006.
   9 ####
  10 #### Authors: Martyn Smith    <martyn@catalyst.net.nz>
  11 ####          Martin Langhoff <martin@catalyst.net.nz>
  12 ####
  13 ####
  14 #### Released under the GNU Public License, version 2.
  15 ####
  16 ####
  17
  18 use strict;
  19 use warnings;
  20 use bytes;
  21
  22 use Fcntl;
  23 use File::Temp qw/tempdir tempfile/;
  24 use File::Basename;
  25
  26 my $log = GITCVS::log->new();
  27 my $cfg;
  28
  29 my $DATE_LIST = {
  30     Jan => "01",
  31     Feb => "02",
  32     Mar => "03",
  33     Apr => "04",
  34     May => "05",
  35     Jun => "06",
  36     Jul => "07",
  37     Aug => "08",
  38     Sep => "09",
  39     Oct => "10",
  40     Nov => "11",
  41     Dec => "12",
  42 };
  43
  44 # Enable autoflush for STDOUT (otherwise the whole thing falls apart)
  45 $| = 1;
  46
  47 #### Definition and mappings of functions ####
  48
  49 my $methods = {
  50     'Root'            => \&req_Root,
  51     'Valid-responses' => \&req_Validresponses,
  52     'valid-requests'  => \&req_validrequests,
  53     'Directory'       => \&req_Directory,
  54     'Entry'           => \&req_Entry,
  55     'Modified'        => \&req_Modified,
  56     'Unchanged'       => \&req_Unchanged,
  57     'Questionable'    => \&req_Questionable,
  58     'Argument'        => \&req_Argument,
  59     'Argumentx'       => \&req_Argument,
  60     'expand-modules'  => \&req_expandmodules,
  61     'add'             => \&req_add,
  62     'remove'          => \&req_remove,
  63     'co'              => \&req_co,
  64     'update'          => \&req_update,
  65     'ci'              => \&req_ci,
  66     'diff'            => \&req_diff,
  67     'log'             => \&req_log,
  68     'rlog'            => \&req_log,
  69     'tag'             => \&req_CATCHALL,
  70     'status'          => \&req_status,
  71     'admin'           => \&req_CATCHALL,
  72     'history'         => \&req_CATCHALL,
  73     'watchers'        => \&req_CATCHALL,
  74     'editors'         => \&req_CATCHALL,
  75     'annotate'        => \&req_annotate,
  76     'Global_option'   => \&req_Globaloption,
  77     #'annotate'        => \&req_CATCHALL,
  78 };
  79
  80 ##############################################
  81
  82
  83 # $state holds all the bits of information the clients sends us that could
  84 # potentially be useful when it comes to actually _doing_ something.
  85 my $state = { prependdir => '' };
  86 $log->info("--------------- STARTING -----------------");
  87
  88 my $TEMP_DIR = tempdir( CLEANUP => 1 );
  89 $log->debug("Temporary directory is '$TEMP_DIR'");
  90
  91 # if we are called with a pserver argument,
  92 # deal with the authentication cat before entering the
  93 # main loop
  94 if (@ARGV && $ARGV[0] eq 'pserver') {
  95     my $line = <STDIN>; chomp $line;
  96     unless( $line eq 'BEGIN AUTH REQUEST') {
  97        die "E Do not understand $line - expecting BEGIN AUTH REQUEST\n";
  98     }
  99     $line = <STDIN>; chomp $line;
 100     req_Root('root', $line) # reuse Root
 101        or die "E Invalid root $line \n";
 102     $line = <STDIN>; chomp $line;
 103     unless ($line eq 'anonymous') {
 104        print "E Only anonymous user allowed via pserver\n";
 105        print "I HATE YOU\n";
 106     }
 107     $line = <STDIN>; chomp $line;    # validate the password?
 108     $line = <STDIN>; chomp $line;
 109     unless ($line eq 'END AUTH REQUEST') {
 110        die "E Do not understand $line -- expecting END AUTH REQUEST\n";
 111     }
 112     print "I LOVE YOU\n";
 113     # and now back to our regular programme...
 114 }
 115
 116 # Keep going until the client closes the connection
 117 while (<STDIN>)
 118 {
 119     chomp;
 120
 121     # Check to see if we've seen this method, and call appropriate function.
 122     if ( /^([\w-]+)(?:\s+(.*))?$/ and defined($methods->{$1}) )
 123     {
 124         # use the $methods hash to call the appropriate sub for this command
 125         #$log->info("Method : $1");
 126         &{$methods->{$1}}($1,$2);
 127     } else {
 128         # log fatal because we don't understand this function. If this happens
 129         # we're fairly screwed because we don't know if the client is expecting
 130         # a response. If it is, the client will hang, we'll hang, and the whole
 131         # thing will be custard.
 132         $log->fatal("Don't understand command $_\n");
 133         die("Unknown command $_");
 134     }
 135 }
 136
 137 $log->debug("Processing time : user=" . (times)[0] . " system=" . (times)[1]);
 138 $log->info("--------------- FINISH -----------------");
 139
 140 # Magic catchall method.
 141 #    This is the method that will handle all commands we haven't yet
 142 #    implemented. It simply sends a warning to the log file indicating a
 143 #    command that hasn't been implemented has been invoked.
 144 sub req_CATCHALL
 145 {
 146     my ( $cmd, $data ) = @_;
 147     $log->warn("Unhandled command : req_$cmd : $data");
 148 }
 149
 150
 151 # Root pathname \n
 152 #     Response expected: no. Tell the server which CVSROOT to use. Note that
 153 #     pathname is a local directory and not a fully qualified CVSROOT variable.
 154 #     pathname must already exist; if creating a new root, use the init
 155 #     request, not Root. pathname does not include the hostname of the server,
 156 #     how to access the server, etc.; by the time the CVS protocol is in use,
 157 #     connection, authentication, etc., are already taken care of. The Root
 158 #     request must be sent only once, and it must be sent before any requests
 159 #     other than Valid-responses, valid-requests, UseUnchanged, Set or init.
 160 sub req_Root
 161 {
 162     my ( $cmd, $data ) = @_;
 163     $log->debug("req_Root : $data");
 164
 165     $state->{CVSROOT} = $data;
 166
 167     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
 168     unless (-d $ENV{GIT_DIR} && -e $ENV{GIT_DIR}.'HEAD') {
 169        print "E $ENV{GIT_DIR} does not seem to be a valid GIT repository\n";
 170         print "E \n";
 171         print "error 1 $ENV{GIT_DIR} is not a valid repository\n";
 172        return 0;
 173     }
 174
 175     my @gitvars = `git-repo-config -l`;
 176     if ($?) {
 177        print "E problems executing git-repo-config on the server -- this is not a git repository or the PATH is not set correctly.\n";
 178         print "E \n";
 179         print "error 1 - problem executing git-repo-config\n";
 180        return 0;
 181     }
 182     foreach my $line ( @gitvars )
 183     {
 184         next unless ( $line =~ /^(.*?)\.(.*?)=(.*)$/ );
 185         $cfg->{$1}{$2} = $3;
 186     }
 187
 188     unless ( defined ( $cfg->{gitcvs}{enabled} ) and $cfg->{gitcvs}{enabled} =~ /^\s*(1|true|yes)\s*$/i )
 189     {
 190         print "E GITCVS emulation needs to be enabled on this repo\n";
 191         print "E the repo config file needs a [gitcvs] section added, and the parameter 'enabled' set to 1\n";
 192         print "E \n";
 193         print "error 1 GITCVS emulation disabled\n";
 194         return 0;
 195     }
 196
 197     if ( defined ( $cfg->{gitcvs}{logfile} ) )
 198     {
 199         $log->setfile($cfg->{gitcvs}{logfile});
 200     } else {
 201         $log->nofile();
 202     }
 203
 204     return 1;
 205 }
 206
 207 # Global_option option \n
 208 #     Response expected: no. Transmit one of the global options `-q', `-Q',
 209 #     `-l', `-t', `-r', or `-n'. option must be one of those strings, no
 210 #     variations (such as combining of options) are allowed. For graceful
 211 #     handling of valid-requests, it is probably better to make new global
 212 #     options separate requests, rather than trying to add them to this
 213 #     request.
 214 sub req_Globaloption
 215 {
 216     my ( $cmd, $data ) = @_;
 217     $log->debug("req_Globaloption : $data");
 218     $state->{globaloptions}{$data} = 1;
 219 }
 220
 221 # Valid-responses request-list \n
 222 #     Response expected: no. Tell the server what responses the client will
 223 #     accept. request-list is a space separated list of tokens.
 224 sub req_Validresponses
 225 {
 226     my ( $cmd, $data ) = @_;
 227     $log->debug("req_Validresponses : $data");
 228
 229     # TODO : re-enable this, currently it's not particularly useful
 230     #$state->{validresponses} = [ split /\s+/, $data ];
 231 }
 232
 233 # valid-requests \n
 234 #     Response expected: yes. Ask the server to send back a Valid-requests
 235 #     response.
 236 sub req_validrequests
 237 {
 238     my ( $cmd, $data ) = @_;
 239
 240     $log->debug("req_validrequests");
 241
 242     $log->debug("SEND : Valid-requests " . join(" ",keys %$methods));
 243     $log->debug("SEND : ok");
 244
 245     print "Valid-requests " . join(" ",keys %$methods) . "\n";
 246     print "ok\n";
 247 }
 248
 249 # Directory local-directory \n
 250 #     Additional data: repository \n. Response expected: no. Tell the server
 251 #     what directory to use. The repository should be a directory name from a
 252 #     previous server response. Note that this both gives a default for Entry
 253 #     and Modified and also for ci and the other commands; normal usage is to
 254 #     send Directory for each directory in which there will be an Entry or
 255 #     Modified, and then a final Directory for the original directory, then the
 256 #     command. The local-directory is relative to the top level at which the
 257 #     command is occurring (i.e. the last Directory which is sent before the
 258 #     command); to indicate that top level, `.' should be sent for
 259 #     local-directory.
 260 sub req_Directory
 261 {
 262     my ( $cmd, $data ) = @_;
 263
 264     my $repository = <STDIN>;
 265     chomp $repository;
 266
 267
 268     $state->{localdir} = $data;
 269     $state->{repository} = $repository;
 270     $state->{path} = $repository;
 271     $state->{path} =~ s/^$state->{CVSROOT}\///;
 272     $state->{module} = $1 if ($state->{path} =~ s/^(.*?)(\/|$)//);
 273     $state->{path} .= "/" if ( $state->{path} =~ /\S/ );
 274
 275     $state->{directory} = $state->{localdir};
 276     $state->{directory} = "" if ( $state->{directory} eq "." );
 277     $state->{directory} .= "/" if ( $state->{directory} =~ /\S/ );
 278
 279     if ( (not defined($state->{prependdir}) or $state->{prependdir} eq '') and $state->{localdir} eq "." and $state->{path} =~ /\S/ )
 280     {
 281         $log->info("Setting prepend to '$state->{path}'");
 282         $state->{prependdir} = $state->{path};
 283         foreach my $entry ( keys %{$state->{entries}} )
 284         {
 285             $state->{entries}{$state->{prependdir} . $entry} = $state->{entries}{$entry};
 286             delete $state->{entries}{$entry};
 287         }
 288     }
 289
 290     if ( defined ( $state->{prependdir} ) )
 291     {
 292         $log->debug("Prepending '$state->{prependdir}' to state|directory");
 293         $state->{directory} = $state->{prependdir} . $state->{directory}
 294     }
 295     $log->debug("req_Directory : localdir=$data repository=$repository path=$state->{path} directory=$state->{directory} module=$state->{module}");
 296 }
 297
 298 # Entry entry-line \n
 299 #     Response expected: no. Tell the server what version of a file is on the
 300 #     local machine. The name in entry-line is a name relative to the directory
 301 #     most recently specified with Directory. If the user is operating on only
 302 #     some files in a directory, Entry requests for only those files need be
 303 #     included. If an Entry request is sent without Modified, Is-modified, or
 304 #     Unchanged, it means the file is lost (does not exist in the working
 305 #     directory). If both Entry and one of Modified, Is-modified, or Unchanged
 306 #     are sent for the same file, Entry must be sent first. For a given file,
 307 #     one can send Modified, Is-modified, or Unchanged, but not more than one
 308 #     of these three.
 309 sub req_Entry
 310 {
 311     my ( $cmd, $data ) = @_;
 312
 313     #$log->debug("req_Entry : $data");
 314
 315     my @data = split(/\//, $data);
 316
 317     $state->{entries}{$state->{directory}.$data[1]} = {
 318         revision    => $data[2],
 319         conflict    => $data[3],
 320         options     => $data[4],
 321         tag_or_date => $data[5],
 322     };
 323
 324     $log->info("Received entry line '$data' => '" . $state->{directory} . $data[1] . "'");
 325 }
 326
 327 # Questionable filename \n
 328 #     Response expected: no. Additional data: no. Tell the server to check
 329 #     whether filename should be ignored, and if not, next time the server
 330 #     sends responses, send (in a M response) `?' followed by the directory and
 331 #     filename. filename must not contain `/'; it needs to be a file in the
 332 #     directory named by the most recent Directory request.
 333 sub req_Questionable
 334 {
 335     my ( $cmd, $data ) = @_;
 336
 337     $log->debug("req_Questionable : $data");
 338     $state->{entries}{$state->{directory}.$data}{questionable} = 1;
 339 }
 340
 341 # add \n
 342 #     Response expected: yes. Add a file or directory. This uses any previous
 343 #     Argument, Directory, Entry, or Modified requests, if they have been sent.
 344 #     The last Directory sent specifies the working directory at the time of
 345 #     the operation. To add a directory, send the directory to be added using
 346 #     Directory and Argument requests.
 347 sub req_add
 348 {
 349     my ( $cmd, $data ) = @_;
 350
 351     argsplit("add");
 352
 353     my $addcount = 0;
 354
 355     foreach my $filename ( @{$state->{args}} )
 356     {
 357         $filename = filecleanup($filename);
 358
 359         unless ( defined ( $state->{entries}{$filename}{modified_filename} ) )
 360         {
 361             print "E cvs add: nothing known about `$filename'\n";
 362             next;
 363         }
 364         # TODO : check we're not squashing an already existing file
 365         if ( defined ( $state->{entries}{$filename}{revision} ) )
 366         {
 367             print "E cvs add: `$filename' has already been entered\n";
 368             next;
 369         }
 370
 371         my ( $filepart, $dirpart ) = filenamesplit($filename, 1);
 372
 373         print "E cvs add: scheduling file `$filename' for addition\n";
 374
 375         print "Checked-in $dirpart\n";
 376         print "$filename\n";
 377         print "/$filepart/0///\n";
 378
 379         $addcount++;
 380     }
 381
 382     if ( $addcount == 1 )
 383     {
 384         print "E cvs add: use `cvs commit' to add this file permanently\n";
 385     }
 386     elsif ( $addcount > 1 )
 387     {
 388         print "E cvs add: use `cvs commit' to add these files permanently\n";
 389     }
 390
 391     print "ok\n";
 392 }
 393
 394 # remove \n
 395 #     Response expected: yes. Remove a file. This uses any previous Argument,
 396 #     Directory, Entry, or Modified requests, if they have been sent. The last
 397 #     Directory sent specifies the working directory at the time of the
 398 #     operation. Note that this request does not actually do anything to the
 399 #     repository; the only effect of a successful remove request is to supply
 400 #     the client with a new entries line containing `-' to indicate a removed
 401 #     file. In fact, the client probably could perform this operation without
 402 #     contacting the server, although using remove may cause the server to
 403 #     perform a few more checks. The client sends a subsequent ci request to
 404 #     actually record the removal in the repository.
 405 sub req_remove
 406 {
 407     my ( $cmd, $data ) = @_;
 408
 409     argsplit("remove");
 410
 411     # Grab a handle to the SQLite db and do any necessary updates
 412     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
 413     $updater->update();
 414
 415     #$log->debug("add state : " . Dumper($state));
 416
 417     my $rmcount = 0;
 418
 419     foreach my $filename ( @{$state->{args}} )
 420     {
 421         $filename = filecleanup($filename);
 422
 423         if ( defined ( $state->{entries}{$filename}{unchanged} ) or defined ( $state->{entries}{$filename}{modified_filename} ) )
 424         {
 425             print "E cvs remove: file `$filename' still in working directory\n";
 426             next;
 427         }
 428
 429         my $meta = $updater->getmeta($filename);
 430         my $wrev = revparse($filename);
 431
 432         unless ( defined ( $wrev ) )
 433         {
 434             print "E cvs remove: nothing known about `$filename'\n";
 435             next;
 436         }
 437
 438         if ( defined($wrev) and $wrev < 0 )
 439         {
 440             print "E cvs remove: file `$filename' already scheduled for removal\n";
 441             next;
 442         }
 443
 444         unless ( $wrev == $meta->{revision} )
 445         {
 446             # TODO : not sure if the format of this message is quite correct.
 447             print "E cvs remove: Up to date check failed for `$filename'\n";
 448             next;
 449         }
 450
 451
 452         my ( $filepart, $dirpart ) = filenamesplit($filename, 1);
 453
 454         print "E cvs remove: scheduling `$filename' for removal\n";
 455
 456         print "Checked-in $dirpart\n";
 457         print "$filename\n";
 458         print "/$filepart/-1.$wrev///\n";
 459
 460         $rmcount++;
 461     }
 462
 463     if ( $rmcount == 1 )
 464     {
 465         print "E cvs remove: use `cvs commit' to remove this file permanently\n";
 466     }
 467     elsif ( $rmcount > 1 )
 468     {
 469         print "E cvs remove: use `cvs commit' to remove these files permanently\n";
 470     }
 471
 472     print "ok\n";
 473 }
 474
 475 # Modified filename \n
 476 #     Response expected: no. Additional data: mode, \n, file transmission. Send
 477 #     the server a copy of one locally modified file. filename is a file within
 478 #     the most recent directory sent with Directory; it must not contain `/'.
 479 #     If the user is operating on only some files in a directory, only those
 480 #     files need to be included. This can also be sent without Entry, if there
 481 #     is no entry for the file.
 482 sub req_Modified
 483 {
 484     my ( $cmd, $data ) = @_;
 485
 486     my $mode = <STDIN>;
 487     chomp $mode;
 488     my $size = <STDIN>;
 489     chomp $size;
 490
 491     # Grab config information
 492     my $blocksize = 8192;
 493     my $bytesleft = $size;
 494     my $tmp;
 495
 496     # Get a filehandle/name to write it to
 497     my ( $fh, $filename ) = tempfile( DIR => $TEMP_DIR );
 498
 499     # Loop over file data writing out to temporary file.
 500     while ( $bytesleft )
 501     {
 502         $blocksize = $bytesleft if ( $bytesleft < $blocksize );
 503         read STDIN, $tmp, $blocksize;
 504         print $fh $tmp;
 505         $bytesleft -= $blocksize;
 506     }
 507
 508     close $fh;
 509
 510     # Ensure we have something sensible for the file mode
 511     if ( $mode =~ /u=(\w+)/ )
 512     {
 513         $mode = $1;
 514     } else {
 515         $mode = "rw";
 516     }
 517
 518     # Save the file data in $state
 519     $state->{entries}{$state->{directory}.$data}{modified_filename} = $filename;
 520     $state->{entries}{$state->{directory}.$data}{modified_mode} = $mode;
 521     $state->{entries}{$state->{directory}.$data}{modified_hash} = `git-hash-object $filename`;
 522     $state->{entries}{$state->{directory}.$data}{modified_hash} =~ s/\s.*$//s;
 523
 524     #$log->debug("req_Modified : file=$data mode=$mode size=$size");
 525 }
 526
 527 # Unchanged filename \n
 528 #     Response expected: no. Tell the server that filename has not been
 529 #     modified in the checked out directory. The filename is a file within the
 530 #     most recent directory sent with Directory; it must not contain `/'.
 531 sub req_Unchanged
 532 {
 533     my ( $cmd, $data ) = @_;
 534
 535     $state->{entries}{$state->{directory}.$data}{unchanged} = 1;
 536
 537     #$log->debug("req_Unchanged : $data");
 538 }
 539
 540 # Argument text \n
 541 #     Response expected: no. Save argument for use in a subsequent command.
 542 #     Arguments accumulate until an argument-using command is given, at which
 543 #     point they are forgotten.
 544 # Argumentx text \n
 545 #     Response expected: no. Append \n followed by text to the current argument
 546 #     being saved.
 547 sub req_Argument
 548 {
 549     my ( $cmd, $data ) = @_;
 550
 551     # Argumentx means: append to last Argument (with a newline in front)
 552
 553     $log->debug("$cmd : $data");
 554
 555     if ( $cmd eq 'Argumentx') {
 556         ${$state->{arguments}}[$#{$state->{arguments}}] .= "\n" . $data;
 557     } else {
 558         push @{$state->{arguments}}, $data;
 559     }
 560 }
 561
 562 # expand-modules \n
 563 #     Response expected: yes. Expand the modules which are specified in the
 564 #     arguments. Returns the data in Module-expansion responses. Note that the
 565 #     server can assume that this is checkout or export, not rtag or rdiff; the
 566 #     latter do not access the working directory and thus have no need to
 567 #     expand modules on the client side. Expand may not be the best word for
 568 #     what this request does. It does not necessarily tell you all the files
 569 #     contained in a module, for example. Basically it is a way of telling you
 570 #     which working directories the server needs to know about in order to
 571 #     handle a checkout of the specified modules. For example, suppose that the
 572 #     server has a module defined by
 573 #   aliasmodule -a 1dir
 574 #     That is, one can check out aliasmodule and it will take 1dir in the
 575 #     repository and check it out to 1dir in the working directory. Now suppose
 576 #     the client already has this module checked out and is planning on using
 577 #     the co request to update it. Without using expand-modules, the client
 578 #     would have two bad choices: it could either send information about all
 579 #     working directories under the current directory, which could be
 580 #     unnecessarily slow, or it could be ignorant of the fact that aliasmodule
 581 #     stands for 1dir, and neglect to send information for 1dir, which would
 582 #     lead to incorrect operation. With expand-modules, the client would first
 583 #     ask for the module to be expanded:
 584 sub req_expandmodules
 585 {
 586     my ( $cmd, $data ) = @_;
 587
 588     argsplit();
 589
 590     $log->debug("req_expandmodules : " . ( defined($data) ? $data : "[NULL]" ) );
 591
 592     unless ( ref $state->{arguments} eq "ARRAY" )
 593     {
 594         print "ok\n";
 595         return;
 596     }
 597
 598     foreach my $module ( @{$state->{arguments}} )
 599     {
 600         $log->debug("SEND : Module-expansion $module");
 601         print "Module-expansion $module\n";
 602     }
 603
 604     print "ok\n";
 605     statecleanup();
 606 }
 607
 608 # co \n
 609 #     Response expected: yes. Get files from the repository. This uses any
 610 #     previous Argument, Directory, Entry, or Modified requests, if they have
 611 #     been sent. Arguments to this command are module names; the client cannot
 612 #     know what directories they correspond to except by (1) just sending the
 613 #     co request, and then seeing what directory names the server sends back in
 614 #     its responses, and (2) the expand-modules request.
 615 sub req_co
 616 {
 617     my ( $cmd, $data ) = @_;
 618
 619     argsplit("co");
 620
 621     my $module = $state->{args}[0];
 622     my $checkout_path = $module;
 623
 624     # use the user specified directory if we're given it
 625     $checkout_path = $state->{opt}{d} if ( exists ( $state->{opt}{d} ) );
 626
 627     $log->debug("req_co : " . ( defined($data) ? $data : "[NULL]" ) );
 628
 629     $log->info("Checking out module '$module' ($state->{CVSROOT}) to '$checkout_path'");
 630
 631     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
 632
 633     # Grab a handle to the SQLite db and do any necessary updates
 634     my $updater = GITCVS::updater->new($state->{CVSROOT}, $module, $log);
 635     $updater->update();
 636
 637     $checkout_path =~ s|/$||; # get rid of trailing slashes
 638
 639     # Eclipse seems to need the Clear-sticky command
 640     # to prepare the 'Entries' file for the new directory.
 641     print "Clear-sticky $checkout_path/\n";
 642     print $state->{CVSROOT} . "/$module/\n";
 643     print "Clear-static-directory $checkout_path/\n";
 644     print $state->{CVSROOT} . "/$module/\n";
 645     print "Clear-sticky $checkout_path/\n"; # yes, twice
 646     print $state->{CVSROOT} . "/$module/\n";
 647     print "Template $checkout_path/\n";
 648     print $state->{CVSROOT} . "/$module/\n";
 649     print "0\n";
 650
 651     # instruct the client that we're checking out to $checkout_path
 652     print "E cvs checkout: Updating $checkout_path\n";
 653
 654     my %seendirs = ();
 655     my $lastdir ='';
 656
 657     # recursive
 658     sub prepdir {
 659        my ($dir, $repodir, $remotedir, $seendirs) = @_;
 660        my $parent = dirname($dir);
 661        $dir       =~ s|/+$||;
 662        $repodir   =~ s|/+$||;
 663        $remotedir =~ s|/+$||;
 664        $parent    =~ s|/+$||;
 665        $log->debug("announcedir $dir, $repodir, $remotedir" );
 666
 667        if ($parent eq '.' || $parent eq './') {
 668            $parent = '';
 669        }
 670        # recurse to announce unseen parents first
 671        if (length($parent) && !exists($seendirs->{$parent})) {
 672            prepdir($parent, $repodir, $remotedir, $seendirs);
 673        }
 674        # Announce that we are going to modify at the parent level
 675        if ($parent) {
 676            print "E cvs checkout: Updating $remotedir/$parent\n";
 677        } else {
 678            print "E cvs checkout: Updating $remotedir\n";
 679        }
 680        print "Clear-sticky $remotedir/$parent/\n";
 681        print "$repodir/$parent/\n";
 682
 683        print "Clear-static-directory $remotedir/$dir/\n";
 684        print "$repodir/$dir/\n";
 685        print "Clear-sticky $remotedir/$parent/\n"; # yes, twice
 686        print "$repodir/$parent/\n";
 687        print "Template $remotedir/$dir/\n";
 688        print "$repodir/$dir/\n";
 689        print "0\n";
 690
 691        $seendirs->{$dir} = 1;
 692     }
 693
 694     foreach my $git ( @{$updater->gethead} )
 695     {
 696         # Don't want to check out deleted files
 697         next if ( $git->{filehash} eq "deleted" );
 698
 699         ( $git->{name}, $git->{dir} ) = filenamesplit($git->{name});
 700
 701        if (length($git->{dir}) && $git->{dir} ne './'
 702            && $git->{dir} ne $lastdir ) {
 703            unless (exists($seendirs{$git->{dir}})) {
 704                prepdir($git->{dir}, $state->{CVSROOT} . "/$module/",
 705                        $checkout_path, \%seendirs);
 706                $lastdir = $git->{dir};
 707                $seendirs{$git->{dir}} = 1;
 708            }
 709            print "E cvs checkout: Updating /$checkout_path/$git->{dir}\n";
 710        }
 711
 712         # modification time of this file
 713         print "Mod-time $git->{modified}\n";
 714
 715         # print some information to the client
 716         if ( defined ( $git->{dir} ) and $git->{dir} ne "./" )
 717         {
 718             print "M U $checkout_path/$git->{dir}$git->{name}\n";
 719         } else {
 720             print "M U $checkout_path/$git->{name}\n";
 721         }
 722
 723        # instruct client we're sending a file to put in this path
 724        print "Created $checkout_path/" . ( defined ( $git->{dir} ) and $git->{dir} ne "./" ? $git->{dir} . "/" : "" ) . "\n";
 725
 726        print $state->{CVSROOT} . "/$module/" . ( defined ( $git->{dir} ) and $git->{dir} ne "./" ? $git->{dir} . "/" : "" ) . "$git->{name}\n";
 727
 728         # this is an "entries" line
 729         print "/$git->{name}/1.$git->{revision}///\n";
 730         # permissions
 731         print "u=$git->{mode},g=$git->{mode},o=$git->{mode}\n";
 732
 733         # transmit file
 734         transmitfile($git->{filehash});
 735     }
 736
 737     print "ok\n";
 738
 739     statecleanup();
 740 }
 741
 742 # update \n
 743 #     Response expected: yes. Actually do a cvs update command. This uses any
 744 #     previous Argument, Directory, Entry, or Modified requests, if they have
 745 #     been sent. The last Directory sent specifies the working directory at the
 746 #     time of the operation. The -I option is not used--files which the client
 747 #     can decide whether to ignore are not mentioned and the client sends the
 748 #     Questionable request for others.
 749 sub req_update
 750 {
 751     my ( $cmd, $data ) = @_;
 752
 753     $log->debug("req_update : " . ( defined($data) ? $data : "[NULL]" ));
 754
 755     argsplit("update");
 756
 757     #
 758     # It may just be a client exploring the available heads/modules
 759     # in that case, list them as top level directories and leave it
 760     # at that. Eclipse uses this technique to offer you a list of
 761     # projects (heads in this case) to checkout.
 762     #
 763     if ($state->{module} eq '') {
 764         print "E cvs update: Updating .\n";
 765         opendir HEADS, $state->{CVSROOT} . '/refs/heads';
 766         while (my $head = readdir(HEADS)) {
 767             if (-f $state->{CVSROOT} . '/refs/heads/' . $head) {
 768                 print "E cvs update: New directory `$head'\n";
 769             }
 770         }
 771         closedir HEADS;
 772         print "ok\n";
 773         return 1;
 774     }
 775
 776
 777     # Grab a handle to the SQLite db and do any necessary updates
 778     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
 779
 780     $updater->update();
 781
 782     argsfromdir($updater);
 783
 784     #$log->debug("update state : " . Dumper($state));
 785
 786     # foreach file specified on the command line ...
 787     foreach my $filename ( @{$state->{args}} )
 788     {
 789         $filename = filecleanup($filename);
 790
 791         $log->debug("Processing file $filename");
 792
 793         # if we have a -C we should pretend we never saw modified stuff
 794         if ( exists ( $state->{opt}{C} ) )
 795         {
 796             delete $state->{entries}{$filename}{modified_hash};
 797             delete $state->{entries}{$filename}{modified_filename};
 798             $state->{entries}{$filename}{unchanged} = 1;
 799         }
 800
 801         my $meta;
 802         if ( defined($state->{opt}{r}) and $state->{opt}{r} =~ /^1\.(\d+)/ )
 803         {
 804             $meta = $updater->getmeta($filename, $1);
 805         } else {
 806             $meta = $updater->getmeta($filename);
 807         }
 808
 809         if ( ! defined $meta )
 810         {
 811             $meta = {
 812                 name => $filename,
 813                 revision => 0,
 814                 filehash => 'added'
 815             };
 816         }
 817
 818         my $oldmeta = $meta;
 819
 820         my $wrev = revparse($filename);
 821
 822         # If the working copy is an old revision, lets get that version too for comparison.
 823         if ( defined($wrev) and $wrev != $meta->{revision} )
 824         {
 825             $oldmeta = $updater->getmeta($filename, $wrev);
 826         }
 827
 828         #$log->debug("Target revision is $meta->{revision}, current working revision is $wrev");
 829
 830         # Files are up to date if the working copy and repo copy have the same revision,
 831         # and the working copy is unmodified _and_ the user hasn't specified -C
 832         next if ( defined ( $wrev )
 833                   and defined($meta->{revision})
 834                   and $wrev == $meta->{revision}
 835                   and $state->{entries}{$filename}{unchanged}
 836                   and not exists ( $state->{opt}{C} ) );
 837
 838         # If the working copy and repo copy have the same revision,
 839         # but the working copy is modified, tell the client it's modified
 840         if ( defined ( $wrev )
 841              and defined($meta->{revision})
 842              and $wrev == $meta->{revision}
 843              and not exists ( $state->{opt}{C} ) )
 844         {
 845             $log->info("Tell the client the file is modified");
 846             print "MT text M \n";
 847             print "MT fname $filename\n";
 848             print "MT newline\n";
 849             next;
 850         }
 851
 852         if ( $meta->{filehash} eq "deleted" )
 853         {
 854             my ( $filepart, $dirpart ) = filenamesplit($filename,1);
 855
 856             $log->info("Removing '$filename' from working copy (no longer in the repo)");
 857
 858             print "E cvs update: `$filename' is no longer in the repository\n";
 859             # Don't want to actually _DO_ the update if -n specified
 860             unless ( $state->{globaloptions}{-n} ) {
 861                 print "Removed $dirpart\n";
 862                 print "$filepart\n";
 863             }
 864         }
 865         elsif ( not defined ( $state->{entries}{$filename}{modified_hash} )
 866                 or $state->{entries}{$filename}{modified_hash} eq $oldmeta->{filehash}
 867                 or $meta->{filehash} eq 'added' )
 868         {
 869             # normal update, just send the new revision (either U=Update,
 870             # or A=Add, or R=Remove)
 871             if ( defined($wrev) && $wrev < 0 )
 872             {
 873                 $log->info("Tell the client the file is scheduled for removal");
 874                 print "MT text R \n";
 875                 print "MT fname $filename\n";
 876                 print "MT newline\n";
 877                 next;
 878             }
 879             elsif ( (!defined($wrev) || $wrev == 0) && (!defined($meta->{revision}) || $meta->{revision} == 0) )
 880             {
 881                 $log->info("Tell the client the file is scheduled for addition");
 882                 print "MT text A \n";
 883                 print "MT fname $filename\n";
 884                 print "MT newline\n";
 885                 next;
 886
 887             }
 888             else {
 889                 $log->info("Updating '$filename' to ".$meta->{revision});
 890                 print "MT +updated\n";
 891                 print "MT text U \n";
 892                 print "MT fname $filename\n";
 893                 print "MT newline\n";
 894                 print "MT -updated\n";
 895             }
 896
 897             my ( $filepart, $dirpart ) = filenamesplit($filename,1);
 898
 899             # Don't want to actually _DO_ the update if -n specified
 900             unless ( $state->{globaloptions}{-n} )
 901             {
 902                 if ( defined ( $wrev ) )
 903                 {
 904                     # instruct client we're sending a file to put in this path as a replacement
 905                     print "Update-existing $dirpart\n";
 906                     $log->debug("Updating existing file 'Update-existing $dirpart'");
 907                 } else {
 908                     # instruct client we're sending a file to put in this path as a new file
 909                     print "Clear-static-directory $dirpart\n";
 910                     print $state->{CVSROOT} . "/$state->{module}/$dirpart\n";
 911                     print "Clear-sticky $dirpart\n";
 912                     print $state->{CVSROOT} . "/$state->{module}/$dirpart\n";
 913
 914                     $log->debug("Creating new file 'Created $dirpart'");
 915                     print "Created $dirpart\n";
 916                 }
 917                 print $state->{CVSROOT} . "/$state->{module}/$filename\n";
 918
 919                 # this is an "entries" line
 920                 $log->debug("/$filepart/1.$meta->{revision}///");
 921                 print "/$filepart/1.$meta->{revision}///\n";
 922
 923                 # permissions
 924                 $log->debug("SEND : u=$meta->{mode},g=$meta->{mode},o=$meta->{mode}");
 925                 print "u=$meta->{mode},g=$meta->{mode},o=$meta->{mode}\n";
 926
 927                 # transmit file
 928                 transmitfile($meta->{filehash});
 929             }
 930         } else {
 931             $log->info("Updating '$filename'");
 932             my ( $filepart, $dirpart ) = filenamesplit($meta->{name},1);
 933
 934             my $dir = tempdir( DIR => $TEMP_DIR, CLEANUP => 1 ) . "/";
 935
 936             chdir $dir;
 937             my $file_local = $filepart . ".mine";
 938             system("ln","-s",$state->{entries}{$filename}{modified_filename}, $file_local);
 939             my $file_old = $filepart . "." . $oldmeta->{revision};
 940             transmitfile($oldmeta->{filehash}, $file_old);
 941             my $file_new = $filepart . "." . $meta->{revision};
 942             transmitfile($meta->{filehash}, $file_new);
 943
 944             # we need to merge with the local changes ( M=successful merge, C=conflict merge )
 945             $log->info("Merging $file_local, $file_old, $file_new");
 946
 947             $log->debug("Temporary directory for merge is $dir");
 948
 949             my $return = system("git", "merge-file", $file_local, $file_old, $file_new);
 950             $return >>= 8;
 951
 952             if ( $return == 0 )
 953             {
 954                 $log->info("Merged successfully");
 955                 print "M M $filename\n";
 956                 $log->debug("Update-existing $dirpart");
 957
 958                 # Don't want to actually _DO_ the update if -n specified
 959                 unless ( $state->{globaloptions}{-n} )
 960                 {
 961                     print "Update-existing $dirpart\n";
 962                     $log->debug($state->{CVSROOT} . "/$state->{module}/$filename");
 963                     print $state->{CVSROOT} . "/$state->{module}/$filename\n";
 964                     $log->debug("/$filepart/1.$meta->{revision}///");
 965                     print "/$filepart/1.$meta->{revision}///\n";
 966                 }
 967             }
 968             elsif ( $return == 1 )
 969             {
 970                 $log->info("Merged with conflicts");
 971                 print "M C $filename\n";
 972
 973                 # Don't want to actually _DO_ the update if -n specified
 974                 unless ( $state->{globaloptions}{-n} )
 975                 {
 976                     print "Update-existing $dirpart\n";
 977                     print $state->{CVSROOT} . "/$state->{module}/$filename\n";
 978                     print "/$filepart/1.$meta->{revision}/+//\n";
 979                 }
 980             }
 981             else
 982             {
 983                 $log->warn("Merge failed");
 984                 next;
 985             }
 986
 987             # Don't want to actually _DO_ the update if -n specified
 988             unless ( $state->{globaloptions}{-n} )
 989             {
 990                 # permissions
 991                 $log->debug("SEND : u=$meta->{mode},g=$meta->{mode},o=$meta->{mode}");
 992                 print "u=$meta->{mode},g=$meta->{mode},o=$meta->{mode}\n";
 993
 994                 # transmit file, format is single integer on a line by itself (file
 995                 # size) followed by the file contents
 996                 # TODO : we should copy files in blocks
 997                 my $data = `cat $file_local`;
 998                 $log->debug("File size : " . length($data));
 999                 print length($data) . "\n";
1000                 print $data;
1001             }
1002
1003             chdir "/";
1004         }
1005
1006     }
1007
1008     print "ok\n";
1009 }
1010
1011 sub req_ci
1012 {
1013     my ( $cmd, $data ) = @_;
1014
1015     argsplit("ci");
1016
1017     #$log->debug("State : " . Dumper($state));
1018
1019     $log->info("req_ci : " . ( defined($data) ? $data : "[NULL]" ));
1020
1021     if ( @ARGV && $ARGV[0] eq 'pserver')
1022     {
1023         print "error 1 pserver access cannot commit\n";
1024         exit;
1025     }
1026
1027     if ( -e $state->{CVSROOT} . "/index" )
1028     {
1029         $log->warn("file 'index' already exists in the git repository");
1030         print "error 1 Index already exists in git repo\n";
1031         exit;
1032     }
1033
1034     my $lockfile = "$state->{CVSROOT}/refs/heads/$state->{module}.lock";
1035     unless ( sysopen(LOCKFILE,$lockfile,O_EXCL|O_CREAT|O_WRONLY) )
1036     {
1037         $log->warn("lockfile '$lockfile' already exists, please try again");
1038         print "error 1 Lock file '$lockfile' already exists, please try again\n";
1039         exit;
1040     }
1041
1042     # Grab a handle to the SQLite db and do any necessary updates
1043     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1044     $updater->update();
1045
1046     my $tmpdir = tempdir ( DIR => $TEMP_DIR );
1047     my ( undef, $file_index ) = tempfile ( DIR => $TEMP_DIR, OPEN => 0 );
1048     $log->info("Lock successful, basing commit on '$tmpdir', index file is '$file_index'");
1049
1050     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
1051     $ENV{GIT_INDEX_FILE} = $file_index;
1052
1053     chdir $tmpdir;
1054
1055     # populate the temporary index based
1056     system("git-read-tree", $state->{module});
1057     unless ($? == 0)
1058     {
1059         die "Error running git-read-tree $state->{module} $file_index $!";
1060     }
1061     $log->info("Created index '$file_index' with for head $state->{module} - exit status $?");
1062
1063
1064     my @committedfiles = ();
1065
1066     # foreach file specified on the command line ...
1067     foreach my $filename ( @{$state->{args}} )
1068     {
1069         my $committedfile = $filename;
1070         $filename = filecleanup($filename);
1071
1072         next unless ( exists $state->{entries}{$filename}{modified_filename} or not $state->{entries}{$filename}{unchanged} );
1073
1074         my $meta = $updater->getmeta($filename);
1075
1076         my $wrev = revparse($filename);
1077
1078         my ( $filepart, $dirpart ) = filenamesplit($filename);
1079
1080         # do a checkout of the file if it part of this tree
1081         if ($wrev) {
1082             system('git-checkout-index', '-f', '-u', $filename);
1083             unless ($? == 0) {
1084                 die "Error running git-checkout-index -f -u $filename : $!";
1085             }
1086         }
1087
1088         my $addflag = 0;
1089         my $rmflag = 0;
1090         $rmflag = 1 if ( defined($wrev) and $wrev < 0 );
1091         $addflag = 1 unless ( -e $filename );
1092
1093         # Do up to date checking
1094         unless ( $addflag or $wrev == $meta->{revision} or ( $rmflag and -$wrev == $meta->{revision} ) )
1095         {
1096             # fail everything if an up to date check fails
1097             print "error 1 Up to date check failed for $filename\n";
1098             close LOCKFILE;
1099             unlink($lockfile);
1100             chdir "/";
1101             exit;
1102         }
1103
1104         push @committedfiles, $committedfile;
1105         $log->info("Committing $filename");
1106
1107         system("mkdir","-p",$dirpart) unless ( -d $dirpart );
1108
1109         unless ( $rmflag )
1110         {
1111             $log->debug("rename $state->{entries}{$filename}{modified_filename} $filename");
1112             rename $state->{entries}{$filename}{modified_filename},$filename;
1113
1114             # Calculate modes to remove
1115             my $invmode = "";
1116             foreach ( qw (r w x) ) { $invmode .= $_ unless ( $state->{entries}{$filename}{modified_mode} =~ /$_/ ); }
1117
1118             $log->debug("chmod u+" . $state->{entries}{$filename}{modified_mode} . "-" . $invmode . " $filename");
1119             system("chmod","u+" .  $state->{entries}{$filename}{modified_mode} . "-" . $invmode, $filename);
1120         }
1121
1122         if ( $rmflag )
1123         {
1124             $log->info("Removing file '$filename'");
1125             unlink($filename);
1126             system("git-update-index", "--remove", $filename);
1127         }
1128         elsif ( $addflag )
1129         {
1130             $log->info("Adding file '$filename'");
1131             system("git-update-index", "--add", $filename);
1132         } else {
1133             $log->info("Updating file '$filename'");
1134             system("git-update-index", $filename);
1135         }
1136     }
1137
1138     unless ( scalar(@committedfiles) > 0 )
1139     {
1140         print "E No files to commit\n";
1141         print "ok\n";
1142         close LOCKFILE;
1143         unlink($lockfile);
1144         chdir "/";
1145         return;
1146     }
1147
1148     my $treehash = `git-write-tree`;
1149     my $parenthash = `cat $ENV{GIT_DIR}refs/heads/$state->{module}`;
1150     chomp $treehash;
1151     chomp $parenthash;
1152
1153     $log->debug("Treehash : $treehash, Parenthash : $parenthash");
1154
1155     # write our commit message out if we have one ...
1156     my ( $msg_fh, $msg_filename ) = tempfile( DIR => $TEMP_DIR );
1157     print $msg_fh $state->{opt}{m};# if ( exists ( $state->{opt}{m} ) );
1158     print $msg_fh "\n\nvia git-CVS emulator\n";
1159     close $msg_fh;
1160
1161     my $commithash = `git-commit-tree $treehash -p $parenthash < $msg_filename`;
1162     $log->info("Commit hash : $commithash");
1163
1164     unless ( $commithash =~ /[a-zA-Z0-9]{40}/ )
1165     {
1166         $log->warn("Commit failed (Invalid commit hash)");
1167         print "error 1 Commit failed (unknown reason)\n";
1168         close LOCKFILE;
1169         unlink($lockfile);
1170         chdir "/";
1171         exit;
1172     }
1173
1174     print LOCKFILE $commithash;
1175
1176     $updater->update();
1177
1178     # foreach file specified on the command line ...
1179     foreach my $filename ( @committedfiles )
1180     {
1181         $filename = filecleanup($filename);
1182
1183         my $meta = $updater->getmeta($filename);
1184         unless (defined $meta->{revision}) {
1185           $meta->{revision} = 1;
1186         }
1187
1188         my ( $filepart, $dirpart ) = filenamesplit($filename, 1);
1189
1190         $log->debug("Checked-in $dirpart : $filename");
1191
1192         if ( defined $meta->{filehash} && $meta->{filehash} eq "deleted" )
1193         {
1194             print "Remove-entry $dirpart\n";
1195             print "$filename\n";
1196         } else {
1197             print "Checked-in $dirpart\n";
1198             print "$filename\n";
1199             print "/$filepart/1.$meta->{revision}///\n";
1200         }
1201     }
1202
1203     close LOCKFILE;
1204     my $reffile = "$ENV{GIT_DIR}refs/heads/$state->{module}";
1205     unlink($reffile);
1206     rename($lockfile, $reffile);
1207     chdir "/";
1208
1209     print "ok\n";
1210 }
1211
1212 sub req_status
1213 {
1214     my ( $cmd, $data ) = @_;
1215
1216     argsplit("status");
1217
1218     $log->info("req_status : " . ( defined($data) ? $data : "[NULL]" ));
1219     #$log->debug("status state : " . Dumper($state));
1220
1221     # Grab a handle to the SQLite db and do any necessary updates
1222     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1223     $updater->update();
1224
1225     # if no files were specified, we need to work out what files we should be providing status on ...
1226     argsfromdir($updater);
1227
1228     # foreach file specified on the command line ...
1229     foreach my $filename ( @{$state->{args}} )
1230     {
1231         $filename = filecleanup($filename);
1232
1233         my $meta = $updater->getmeta($filename);
1234         my $oldmeta = $meta;
1235
1236         my $wrev = revparse($filename);
1237
1238         # If the working copy is an old revision, lets get that version too for comparison.
1239         if ( defined($wrev) and $wrev != $meta->{revision} )
1240         {
1241             $oldmeta = $updater->getmeta($filename, $wrev);
1242         }
1243
1244         # TODO : All possible statuses aren't yet implemented
1245         my $status;
1246         # Files are up to date if the working copy and repo copy have the same revision, and the working copy is unmodified
1247         $status = "Up-to-date" if ( defined ( $wrev ) and defined($meta->{revision}) and $wrev == $meta->{revision}
1248                                     and
1249                                     ( ( $state->{entries}{$filename}{unchanged} and ( not defined ( $state->{entries}{$filename}{conflict} ) or $state->{entries}{$filename}{conflict} !~ /^\+=/ ) )
1250                                       or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $meta->{filehash} ) )
1251                                    );
1252
1253         # Need checkout if the working copy has an older revision than the repo copy, and the working copy is unmodified
1254         $status ||= "Needs Checkout" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and $meta->{revision} > $wrev
1255                                           and
1256                                           ( $state->{entries}{$filename}{unchanged}
1257                                             or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $oldmeta->{filehash} ) )
1258                                         );
1259
1260         # Need checkout if it exists in the repo but doesn't have a working copy
1261         $status ||= "Needs Checkout" if ( not defined ( $wrev ) and defined ( $meta->{revision} ) );
1262
1263         # Locally modified if working copy and repo copy have the same revision but there are local changes
1264         $status ||= "Locally Modified" if ( defined ( $wrev ) and defined($meta->{revision}) and $wrev == $meta->{revision} and $state->{entries}{$filename}{modified_filename} );
1265
1266         # Needs Merge if working copy revision is less than repo copy and there are local changes
1267         $status ||= "Needs Merge" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and $meta->{revision} > $wrev and $state->{entries}{$filename}{modified_filename} );
1268
1269         $status ||= "Locally Added" if ( defined ( $state->{entries}{$filename}{revision} ) and not defined ( $meta->{revision} ) );
1270         $status ||= "Locally Removed" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and -$wrev == $meta->{revision} );
1271         $status ||= "Unresolved Conflict" if ( defined ( $state->{entries}{$filename}{conflict} ) and $state->{entries}{$filename}{conflict} =~ /^\+=/ );
1272         $status ||= "File had conflicts on merge" if ( 0 );
1273
1274         $status ||= "Unknown";
1275
1276         print "M ===================================================================\n";
1277         print "M File: $filename\tStatus: $status\n";
1278         if ( defined($state->{entries}{$filename}{revision}) )
1279         {
1280             print "M Working revision:\t" . $state->{entries}{$filename}{revision} . "\n";
1281         } else {
1282             print "M Working revision:\tNo entry for $filename\n";
1283         }
1284         if ( defined($meta->{revision}) )
1285         {
1286             print "M Repository revision:\t1." . $meta->{revision} . "\t$state->{repository}/$filename,v\n";
1287             print "M Sticky Tag:\t\t(none)\n";
1288             print "M Sticky Date:\t\t(none)\n";
1289             print "M Sticky Options:\t\t(none)\n";
1290         } else {
1291             print "M Repository revision:\tNo revision control file\n";
1292         }
1293         print "M\n";
1294     }
1295
1296     print "ok\n";
1297 }
1298
1299 sub req_diff
1300 {
1301     my ( $cmd, $data ) = @_;
1302
1303     argsplit("diff");
1304
1305     $log->debug("req_diff : " . ( defined($data) ? $data : "[NULL]" ));
1306     #$log->debug("status state : " . Dumper($state));
1307
1308     my ($revision1, $revision2);
1309     if ( defined ( $state->{opt}{r} ) and ref $state->{opt}{r} eq "ARRAY" )
1310     {
1311         $revision1 = $state->{opt}{r}[0];
1312         $revision2 = $state->{opt}{r}[1];
1313     } else {
1314         $revision1 = $state->{opt}{r};
1315     }
1316
1317     $revision1 =~ s/^1\.// if ( defined ( $revision1 ) );
1318     $revision2 =~ s/^1\.// if ( defined ( $revision2 ) );
1319
1320     $log->debug("Diffing revisions " . ( defined($revision1) ? $revision1 : "[NULL]" ) . " and " . ( defined($revision2) ? $revision2 : "[NULL]" ) );
1321
1322     # Grab a handle to the SQLite db and do any necessary updates
1323     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1324     $updater->update();
1325
1326     # if no files were specified, we need to work out what files we should be providing status on ...
1327     argsfromdir($updater);
1328
1329     # foreach file specified on the command line ...
1330     foreach my $filename ( @{$state->{args}} )
1331     {
1332         $filename = filecleanup($filename);
1333
1334         my ( $fh, $file1, $file2, $meta1, $meta2, $filediff );
1335
1336         my $wrev = revparse($filename);
1337
1338         # We need _something_ to diff against
1339         next unless ( defined ( $wrev ) );
1340
1341         # if we have a -r switch, use it
1342         if ( defined ( $revision1 ) )
1343         {
1344             ( undef, $file1 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1345             $meta1 = $updater->getmeta($filename, $revision1);
1346             unless ( defined ( $meta1 ) and $meta1->{filehash} ne "deleted" )
1347             {
1348                 print "E File $filename at revision 1.$revision1 doesn't exist\n";
1349                 next;
1350             }
1351             transmitfile($meta1->{filehash}, $file1);
1352         }
1353         # otherwise we just use the working copy revision
1354         else
1355         {
1356             ( undef, $file1 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1357             $meta1 = $updater->getmeta($filename, $wrev);
1358             transmitfile($meta1->{filehash}, $file1);
1359         }
1360
1361         # if we have a second -r switch, use it too
1362         if ( defined ( $revision2 ) )
1363         {
1364             ( undef, $file2 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1365             $meta2 = $updater->getmeta($filename, $revision2);
1366
1367             unless ( defined ( $meta2 ) and $meta2->{filehash} ne "deleted" )
1368             {
1369                 print "E File $filename at revision 1.$revision2 doesn't exist\n";
1370                 next;
1371             }
1372
1373             transmitfile($meta2->{filehash}, $file2);
1374         }
1375         # otherwise we just use the working copy
1376         else
1377         {
1378             $file2 = $state->{entries}{$filename}{modified_filename};
1379         }
1380
1381         # if we have been given -r, and we don't have a $file2 yet, lets get one
1382         if ( defined ( $revision1 ) and not defined ( $file2 ) )
1383         {
1384             ( undef, $file2 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1385             $meta2 = $updater->getmeta($filename, $wrev);
1386             transmitfile($meta2->{filehash}, $file2);
1387         }
1388
1389         # We need to have retrieved something useful
1390         next unless ( defined ( $meta1 ) );
1391
1392         # Files to date if the working copy and repo copy have the same revision, and the working copy is unmodified
1393         next if ( not defined ( $meta2 ) and $wrev == $meta1->{revision}
1394                   and
1395                    ( ( $state->{entries}{$filename}{unchanged} and ( not defined ( $state->{entries}{$filename}{conflict} ) or $state->{entries}{$filename}{conflict} !~ /^\+=/ ) )
1396                      or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $meta1->{filehash} ) )
1397                   );
1398
1399         # Apparently we only show diffs for locally modified files
1400         next unless ( defined($meta2) or defined ( $state->{entries}{$filename}{modified_filename} ) );
1401
1402         print "M Index: $filename\n";
1403         print "M ===================================================================\n";
1404         print "M RCS file: $state->{CVSROOT}/$state->{module}/$filename,v\n";
1405         print "M retrieving revision 1.$meta1->{revision}\n" if ( defined ( $meta1 ) );
1406         print "M retrieving revision 1.$meta2->{revision}\n" if ( defined ( $meta2 ) );
1407         print "M diff ";
1408         foreach my $opt ( keys %{$state->{opt}} )
1409         {
1410             if ( ref $state->{opt}{$opt} eq "ARRAY" )
1411             {
1412                 foreach my $value ( @{$state->{opt}{$opt}} )
1413                 {
1414                     print "-$opt $value ";
1415                 }
1416             } else {
1417                 print "-$opt ";
1418                 print "$state->{opt}{$opt} " if ( defined ( $state->{opt}{$opt} ) );
1419             }
1420         }
1421         print "$filename\n";
1422
1423         $log->info("Diffing $filename -r $meta1->{revision} -r " . ( $meta2->{revision} or "workingcopy" ));
1424
1425         ( $fh, $filediff ) = tempfile ( DIR => $TEMP_DIR );
1426
1427         if ( exists $state->{opt}{u} )
1428         {
1429             system("diff -u -L '$filename revision 1.$meta1->{revision}' -L '$filename " . ( defined($meta2->{revision}) ? "revision 1.$meta2->{revision}" : "working copy" ) . "' $file1 $file2 > $filediff");
1430         } else {
1431             system("diff $file1 $file2 > $filediff");
1432         }
1433
1434         while ( <$fh> )
1435         {
1436             print "M $_";
1437         }
1438         close $fh;
1439     }
1440
1441     print "ok\n";
1442 }
1443
1444 sub req_log
1445 {
1446     my ( $cmd, $data ) = @_;
1447
1448     argsplit("log");
1449
1450     $log->debug("req_log : " . ( defined($data) ? $data : "[NULL]" ));
1451     #$log->debug("log state : " . Dumper($state));
1452
1453     my ( $minrev, $maxrev );
1454     if ( defined ( $state->{opt}{r} ) and $state->{opt}{r} =~ /([\d.]+)?(::?)([\d.]+)?/ )
1455     {
1456         my $control = $2;
1457         $minrev = $1;
1458         $maxrev = $3;
1459         $minrev =~ s/^1\.// if ( defined ( $minrev ) );
1460         $maxrev =~ s/^1\.// if ( defined ( $maxrev ) );
1461         $minrev++ if ( defined($minrev) and $control eq "::" );
1462     }
1463
1464     # Grab a handle to the SQLite db and do any necessary updates
1465     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1466     $updater->update();
1467
1468     # if no files were specified, we need to work out what files we should be providing status on ...
1469     argsfromdir($updater);
1470
1471     # foreach file specified on the command line ...
1472     foreach my $filename ( @{$state->{args}} )
1473     {
1474         $filename = filecleanup($filename);
1475
1476         my $headmeta = $updater->getmeta($filename);
1477
1478         my $revisions = $updater->getlog($filename);
1479         my $totalrevisions = scalar(@$revisions);
1480
1481         if ( defined ( $minrev ) )
1482         {
1483             $log->debug("Removing revisions less than $minrev");
1484             while ( scalar(@$revisions) > 0 and $revisions->[-1]{revision} < $minrev )
1485             {
1486                 pop @$revisions;
1487             }
1488         }
1489         if ( defined ( $maxrev ) )
1490         {
1491             $log->debug("Removing revisions greater than $maxrev");
1492             while ( scalar(@$revisions) > 0 and $revisions->[0]{revision} > $maxrev )
1493             {
1494                 shift @$revisions;
1495             }
1496         }
1497
1498         next unless ( scalar(@$revisions) );
1499
1500         print "M \n";
1501         print "M RCS file: $state->{CVSROOT}/$state->{module}/$filename,v\n";
1502         print "M Working file: $filename\n";
1503         print "M head: 1.$headmeta->{revision}\n";
1504         print "M branch:\n";
1505         print "M locks: strict\n";
1506         print "M access list:\n";
1507         print "M symbolic names:\n";
1508         print "M keyword substitution: kv\n";
1509         print "M total revisions: $totalrevisions;\tselected revisions: " . scalar(@$revisions) . "\n";
1510         print "M description:\n";
1511
1512         foreach my $revision ( @$revisions )
1513         {
1514             print "M ----------------------------\n";
1515             print "M revision 1.$revision->{revision}\n";
1516             # reformat the date for log output
1517             $revision->{modified} = sprintf('%04d/%02d/%02d %s', $3, $DATE_LIST->{$2}, $1, $4 ) if ( $revision->{modified} =~ /(\d+)\s+(\w+)\s+(\d+)\s+(\S+)/ and defined($DATE_LIST->{$2}) );
1518             $revision->{author} =~ s/\s+.*//;
1519             $revision->{author} =~ s/^(.{8}).*/$1/;
1520             print "M date: $revision->{modified};  author: $revision->{author};  state: " . ( $revision->{filehash} eq "deleted" ? "dead" : "Exp" ) . ";  lines: +2 -3\n";
1521             my $commitmessage = $updater->commitmessage($revision->{commithash});
1522             $commitmessage =~ s/^/M /mg;
1523             print $commitmessage . "\n";
1524         }
1525         print "M =============================================================================\n";
1526     }
1527
1528     print "ok\n";
1529 }
1530
1531 sub req_annotate
1532 {
1533     my ( $cmd, $data ) = @_;
1534
1535     argsplit("annotate");
1536
1537     $log->info("req_annotate : " . ( defined($data) ? $data : "[NULL]" ));
1538     #$log->debug("status state : " . Dumper($state));
1539
1540     # Grab a handle to the SQLite db and do any necessary updates
1541     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1542     $updater->update();
1543
1544     # if no files were specified, we need to work out what files we should be providing annotate on ...
1545     argsfromdir($updater);
1546
1547     # we'll need a temporary checkout dir
1548     my $tmpdir = tempdir ( DIR => $TEMP_DIR );
1549     my ( undef, $file_index ) = tempfile ( DIR => $TEMP_DIR, OPEN => 0 );
1550     $log->info("Temp checkoutdir creation successful, basing annotate session work on '$tmpdir', index file is '$file_index'");
1551
1552     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
1553     $ENV{GIT_INDEX_FILE} = $file_index;
1554
1555     chdir $tmpdir;
1556
1557     # foreach file specified on the command line ...
1558     foreach my $filename ( @{$state->{args}} )
1559     {
1560         $filename = filecleanup($filename);
1561
1562         my $meta = $updater->getmeta($filename);
1563
1564         next unless ( $meta->{revision} );
1565
1566         # get all the commits that this file was in
1567         # in dense format -- aka skip dead revisions
1568         my $revisions   = $updater->gethistorydense($filename);
1569         my $lastseenin  = $revisions->[0][2];
1570
1571         # populate the temporary index based on the latest commit were we saw
1572         # the file -- but do it cheaply without checking out any files
1573         # TODO: if we got a revision from the client, use that instead
1574         # to look up the commithash in sqlite (still good to default to
1575         # the current head as we do now)
1576         system("git-read-tree", $lastseenin);
1577         unless ($? == 0)
1578         {
1579             die "Error running git-read-tree $lastseenin $file_index $!";
1580         }
1581         $log->info("Created index '$file_index' with commit $lastseenin - exit status $?");
1582
1583         # do a checkout of the file
1584         system('git-checkout-index', '-f', '-u', $filename);
1585         unless ($? == 0) {
1586             die "Error running git-checkout-index -f -u $filename : $!";
1587         }
1588
1589         $log->info("Annotate $filename");
1590
1591         # Prepare a file with the commits from the linearized
1592         # history that annotate should know about. This prevents
1593         # git-jsannotate telling us about commits we are hiding
1594         # from the client.
1595
1596         open(ANNOTATEHINTS, ">$tmpdir/.annotate_hints") or die "Error opening > $tmpdir/.annotate_hints $!";
1597         for (my $i=0; $i < @$revisions; $i++)
1598         {
1599             print ANNOTATEHINTS $revisions->[$i][2];
1600             if ($i+1 < @$revisions) { # have we got a parent?
1601                 print ANNOTATEHINTS ' ' . $revisions->[$i+1][2];
1602             }
1603             print ANNOTATEHINTS "\n";
1604         }
1605
1606         print ANNOTATEHINTS "\n";
1607         close ANNOTATEHINTS;
1608
1609         my $annotatecmd = 'git-annotate';
1610         open(ANNOTATE, "-|", $annotatecmd, '-l', '-S', "$tmpdir/.annotate_hints", $filename)
1611             or die "Error invoking $annotatecmd -l -S $tmpdir/.annotate_hints $filename : $!";
1612         my $metadata = {};
1613         print "E Annotations for $filename\n";
1614         print "E ***************\n";
1615         while ( <ANNOTATE> )
1616         {
1617             if (m/^([a-zA-Z0-9]{40})\t\([^\)]*\)(.*)$/i)
1618             {
1619                 my $commithash = $1;
1620                 my $data = $2;
1621                 unless ( defined ( $metadata->{$commithash} ) )
1622                 {
1623                     $metadata->{$commithash} = $updater->getmeta($filename, $commithash);
1624                     $metadata->{$commithash}{author} =~ s/\s+.*//;
1625                     $metadata->{$commithash}{author} =~ s/^(.{8}).*/$1/;
1626                     $metadata->{$commithash}{modified} = sprintf("%02d-%s-%02d", $1, $2, $3) if ( $metadata->{$commithash}{modified} =~ /^(\d+)\s(\w+)\s\d\d(\d\d)/ );
1627                 }
1628                 printf("M 1.%-5d      (%-8s %10s): %s\n",
1629                     $metadata->{$commithash}{revision},
1630                     $metadata->{$commithash}{author},
1631                     $metadata->{$commithash}{modified},
1632                     $data
1633                 );
1634             } else {
1635                 $log->warn("Error in annotate output! LINE: $_");
1636                 print "E Annotate error \n";
1637                 next;
1638             }
1639         }
1640         close ANNOTATE;
1641     }
1642
1643     # done; get out of the tempdir
1644     chdir "/";
1645
1646     print "ok\n";
1647
1648 }
1649
1650 # This method takes the state->{arguments} array and produces two new arrays.
1651 # The first is $state->{args} which is everything before the '--' argument, and
1652 # the second is $state->{files} which is everything after it.
1653 sub argsplit
1654 {
1655     return unless( defined($state->{arguments}) and ref $state->{arguments} eq "ARRAY" );
1656
1657     my $type = shift;
1658
1659     $state->{args} = [];
1660     $state->{files} = [];
1661     $state->{opt} = {};
1662
1663     if ( defined($type) )
1664     {
1665         my $opt = {};
1666         $opt = { A => 0, N => 0, P => 0, R => 0, c => 0, f => 0, l => 0, n => 0, p => 0, s => 0, r => 1, D => 1, d => 1, k => 1, j => 1, } if ( $type eq "co" );
1667         $opt = { v => 0, l => 0, R => 0 } if ( $type eq "status" );
1668         $opt = { A => 0, P => 0, C => 0, d => 0, f => 0, l => 0, R => 0, p => 0, k => 1, r => 1, D => 1, j => 1, I => 1, W => 1 } if ( $type eq "update" );
1669         $opt = { l => 0, R => 0, k => 1, D => 1, D => 1, r => 2 } if ( $type eq "diff" );
1670         $opt = { c => 0, R => 0, l => 0, f => 0, F => 1, m => 1, r => 1 } if ( $type eq "ci" );
1671         $opt = { k => 1, m => 1 } if ( $type eq "add" );
1672         $opt = { f => 0, l => 0, R => 0 } if ( $type eq "remove" );
1673         $opt = { l => 0, b => 0, h => 0, R => 0, t => 0, N => 0, S => 0, r => 1, d => 1, s => 1, w => 1 } if ( $type eq "log" );
1674
1675
1676         while ( scalar ( @{$state->{arguments}} ) > 0 )
1677         {
1678             my $arg = shift @{$state->{arguments}};
1679
1680             next if ( $arg eq "--" );
1681             next unless ( $arg =~ /\S/ );
1682
1683             # if the argument looks like a switch
1684             if ( $arg =~ /^-(\w)(.*)/ )
1685             {
1686                 # if it's a switch that takes an argument
1687                 if ( $opt->{$1} )
1688                 {
1689                     # If this switch has already been provided
1690                     if ( $opt->{$1} > 1 and exists ( $state->{opt}{$1} ) )
1691                     {
1692                         $state->{opt}{$1} = [ $state->{opt}{$1} ];
1693                         if ( length($2) > 0 )
1694                         {
1695                             push @{$state->{opt}{$1}},$2;
1696                         } else {
1697                             push @{$state->{opt}{$1}}, shift @{$state->{arguments}};
1698                         }
1699                     } else {
1700                         # if there's extra data in the arg, use that as the argument for the switch
1701                         if ( length($2) > 0 )
1702                         {
1703                             $state->{opt}{$1} = $2;
1704                         } else {
1705                             $state->{opt}{$1} = shift @{$state->{arguments}};
1706                         }
1707                     }
1708                 } else {
1709                     $state->{opt}{$1} = undef;
1710                 }
1711             }
1712             else
1713             {
1714                 push @{$state->{args}}, $arg;
1715             }
1716         }
1717     }
1718     else
1719     {
1720         my $mode = 0;
1721
1722         foreach my $value ( @{$state->{arguments}} )
1723         {
1724             if ( $value eq "--" )
1725             {
1726                 $mode++;
1727                 next;
1728             }
1729             push @{$state->{args}}, $value if ( $mode == 0 );
1730             push @{$state->{files}}, $value if ( $mode == 1 );
1731         }
1732     }
1733 }
1734
1735 # This method uses $state->{directory} to populate $state->{args} with a list of filenames
1736 sub argsfromdir
1737 {
1738     my $updater = shift;
1739
1740     $state->{args} = [] if ( scalar(@{$state->{args}}) == 1 and $state->{args}[0] eq "." );
1741
1742     return if ( scalar ( @{$state->{args}} ) > 1 );
1743
1744     my @gethead = @{$updater->gethead};
1745
1746     # push added files
1747     foreach my $file (keys %{$state->{entries}}) {
1748         if ( exists $state->{entries}{$file}{revision} &&
1749                 $state->{entries}{$file}{revision} == 0 )
1750         {
1751             push @gethead, { name => $file, filehash => 'added' };
1752         }
1753     }
1754
1755     if ( scalar(@{$state->{args}}) == 1 )
1756     {
1757         my $arg = $state->{args}[0];
1758         $arg .= $state->{prependdir} if ( defined ( $state->{prependdir} ) );
1759
1760         $log->info("Only one arg specified, checking for directory expansion on '$arg'");
1761
1762         foreach my $file ( @gethead )
1763         {
1764             next if ( $file->{filehash} eq "deleted" and not defined ( $state->{entries}{$file->{name}} ) );
1765             next unless ( $file->{name} =~ /^$arg\// or $file->{name} eq $arg  );
1766             push @{$state->{args}}, $file->{name};
1767         }
1768
1769         shift @{$state->{args}} if ( scalar(@{$state->{args}}) > 1 );
1770     } else {
1771         $log->info("Only one arg specified, populating file list automatically");
1772
1773         $state->{args} = [];
1774
1775         foreach my $file ( @gethead )
1776         {
1777             next if ( $file->{filehash} eq "deleted" and not defined ( $state->{entries}{$file->{name}} ) );
1778             next unless ( $file->{name} =~ s/^$state->{prependdir}// );
1779             push @{$state->{args}}, $file->{name};
1780         }
1781     }
1782 }
1783
1784 # This method cleans up the $state variable after a command that uses arguments has run
1785 sub statecleanup
1786 {
1787     $state->{files} = [];
1788     $state->{args} = [];
1789     $state->{arguments} = [];
1790     $state->{entries} = {};
1791 }
1792
1793 sub revparse
1794 {
1795     my $filename = shift;
1796
1797     return undef unless ( defined ( $state->{entries}{$filename}{revision} ) );
1798
1799     return $1 if ( $state->{entries}{$filename}{revision} =~ /^1\.(\d+)/ );
1800     return -$1 if ( $state->{entries}{$filename}{revision} =~ /^-1\.(\d+)/ );
1801
1802     return undef;
1803 }
1804
1805 # This method takes a file hash and does a CVS "file transfer" which transmits the
1806 # size of the file, and then the file contents.
1807 # If a second argument $targetfile is given, the file is instead written out to
1808 # a file by the name of $targetfile
1809 sub transmitfile
1810 {
1811     my $filehash = shift;
1812     my $targetfile = shift;
1813
1814     if ( defined ( $filehash ) and $filehash eq "deleted" )
1815     {
1816         $log->warn("filehash is 'deleted'");
1817         return;
1818     }
1819
1820     die "Need filehash" unless ( defined ( $filehash ) and $filehash =~ /^[a-zA-Z0-9]{40}$/ );
1821
1822     my $type = `git-cat-file -t $filehash`;
1823     chomp $type;
1824
1825     die ( "Invalid type '$type' (expected 'blob')" ) unless ( defined ( $type ) and $type eq "blob" );
1826
1827     my $size = `git-cat-file -s $filehash`;
1828     chomp $size;
1829
1830     $log->debug("transmitfile($filehash) size=$size, type=$type");
1831
1832     if ( open my $fh, '-|', "git-cat-file", "blob", $filehash )
1833     {
1834         if ( defined ( $targetfile ) )
1835         {
1836             open NEWFILE, ">", $targetfile or die("Couldn't open '$targetfile' for writing : $!");
1837             print NEWFILE $_ while ( <$fh> );
1838             close NEWFILE;
1839         } else {
1840             print "$size\n";
1841             print while ( <$fh> );
1842         }
1843         close $fh or die ("Couldn't close filehandle for transmitfile()");
1844     } else {
1845         die("Couldn't execute git-cat-file");
1846     }
1847 }
1848
1849 # This method takes a file name, and returns ( $dirpart, $filepart ) which
1850 # refers to the directory portion and the file portion of the filename
1851 # respectively
1852 sub filenamesplit
1853 {
1854     my $filename = shift;
1855     my $fixforlocaldir = shift;
1856
1857     my ( $filepart, $dirpart ) = ( $filename, "." );
1858     ( $filepart, $dirpart ) = ( $2, $1 ) if ( $filename =~ /(.*)\/(.*)/ );
1859     $dirpart .= "/";
1860
1861     if ( $fixforlocaldir )
1862     {
1863         $dirpart =~ s/^$state->{prependdir}//;
1864     }
1865
1866     return ( $filepart, $dirpart );
1867 }
1868
1869 sub filecleanup
1870 {
1871     my $filename = shift;
1872
1873     return undef unless(defined($filename));
1874     if ( $filename =~ /^\// )
1875     {
1876         print "E absolute filenames '$filename' not supported by server\n";
1877         return undef;
1878     }
1879
1880     $filename =~ s/^\.\///g;
1881     $filename = $state->{prependdir} . $filename;
1882     return $filename;
1883 }
1884
1885 package GITCVS::log;
1886
1887 ####
1888 #### Copyright The Open University UK - 2006.
1889 ####
1890 #### Authors: Martyn Smith    <martyn@catalyst.net.nz>
1891 ####          Martin Langhoff <martin@catalyst.net.nz>
1892 ####
1893 ####
1894
1895 use strict;
1896 use warnings;
1897
1898 =head1 NAME
1899
1900 GITCVS::log
1901
1902 =head1 DESCRIPTION
1903
1904 This module provides very crude logging with a similar interface to
1905 Log::Log4perl
1906
1907 =head1 METHODS
1908
1909 =cut
1910
1911 =head2 new
1912
1913 Creates a new log object, optionally you can specify a filename here to
1914 indicate the file to log to. If no log file is specified, you can specify one
1915 later with method setfile, or indicate you no longer want logging with method
1916 nofile.
1917
1918 Until one of these methods is called, all log calls will buffer messages ready
1919 to write out.
1920
1921 =cut
1922 sub new
1923 {
1924     my $class = shift;
1925     my $filename = shift;
1926
1927     my $self = {};
1928
1929     bless $self, $class;
1930
1931     if ( defined ( $filename ) )
1932     {
1933         open $self->{fh}, ">>", $filename or die("Couldn't open '$filename' for writing : $!");
1934     }
1935
1936     return $self;
1937 }
1938
1939 =head2 setfile
1940
1941 This methods takes a filename, and attempts to open that file as the log file.
1942 If successful, all buffered data is written out to the file, and any further
1943 logging is written directly to the file.
1944
1945 =cut
1946 sub setfile
1947 {
1948     my $self = shift;
1949     my $filename = shift;
1950
1951     if ( defined ( $filename ) )
1952     {
1953         open $self->{fh}, ">>", $filename or die("Couldn't open '$filename' for writing : $!");
1954     }
1955
1956     return unless ( defined ( $self->{buffer} ) and ref $self->{buffer} eq "ARRAY" );
1957
1958     while ( my $line = shift @{$self->{buffer}} )
1959     {
1960         print {$self->{fh}} $line;
1961     }
1962 }
1963
1964 =head2 nofile
1965
1966 This method indicates no logging is going to be used. It flushes any entries in
1967 the internal buffer, and sets a flag to ensure no further data is put there.
1968
1969 =cut
1970 sub nofile
1971 {
1972     my $self = shift;
1973
1974     $self->{nolog} = 1;
1975
1976     return unless ( defined ( $self->{buffer} ) and ref $self->{buffer} eq "ARRAY" );
1977
1978     $self->{buffer} = [];
1979 }
1980
1981 =head2 _logopen
1982
1983 Internal method. Returns true if the log file is open, false otherwise.
1984
1985 =cut
1986 sub _logopen
1987 {
1988     my $self = shift;
1989
1990     return 1 if ( defined ( $self->{fh} ) and ref $self->{fh} eq "GLOB" );
1991     return 0;
1992 }
1993
1994 =head2 debug info warn fatal
1995
1996 These four methods are wrappers to _log. They provide the actual interface for
1997 logging data.
1998
1999 =cut
2000 sub debug { my $self = shift; $self->_log("debug", @_); }
2001 sub info  { my $self = shift; $self->_log("info" , @_); }
2002 sub warn  { my $self = shift; $self->_log("warn" , @_); }
2003 sub fatal { my $self = shift; $self->_log("fatal", @_); }
2004
2005 =head2 _log
2006
2007 This is an internal method called by the logging functions. It generates a
2008 timestamp and pushes the logged line either to file, or internal buffer.
2009
2010 =cut
2011 sub _log
2012 {
2013     my $self = shift;
2014     my $level = shift;
2015
2016     return if ( $self->{nolog} );
2017
2018     my @time = localtime;
2019     my $timestring = sprintf("%4d-%02d-%02d %02d:%02d:%02d : %-5s",
2020         $time[5] + 1900,
2021         $time[4] + 1,
2022         $time[3],
2023         $time[2],
2024         $time[1],
2025         $time[0],
2026         uc $level,
2027     );
2028
2029     if ( $self->_logopen )
2030     {
2031         print {$self->{fh}} $timestring . " - " . join(" ",@_) . "\n";
2032     } else {
2033         push @{$self->{buffer}}, $timestring . " - " . join(" ",@_) . "\n";
2034     }
2035 }
2036
2037 =head2 DESTROY
2038
2039 This method simply closes the file handle if one is open
2040
2041 =cut
2042 sub DESTROY
2043 {
2044     my $self = shift;
2045
2046     if ( $self->_logopen )
2047     {
2048         close $self->{fh};
2049     }
2050 }
2051
2052 package GITCVS::updater;
2053
2054 ####
2055 #### Copyright The Open University UK - 2006.
2056 ####
2057 #### Authors: Martyn Smith    <martyn@catalyst.net.nz>
2058 ####          Martin Langhoff <martin@catalyst.net.nz>
2059 ####
2060 ####
2061
2062 use strict;
2063 use warnings;
2064 use DBI;
2065
2066 =head1 METHODS
2067
2068 =cut
2069
2070 =head2 new
2071
2072 =cut
2073 sub new
2074 {
2075     my $class = shift;
2076     my $config = shift;
2077     my $module = shift;
2078     my $log = shift;
2079
2080     die "Need to specify a git repository" unless ( defined($config) and -d $config );
2081     die "Need to specify a module" unless ( defined($module) );
2082
2083     $class = ref($class) || $class;
2084
2085     my $self = {};
2086
2087     bless $self, $class;
2088
2089     $self->{dbdir} = $config . "/";
2090     die "Database dir '$self->{dbdir}' isn't a directory" unless ( defined($self->{dbdir}) and -d $self->{dbdir} );
2091
2092     $self->{module} = $module;
2093     $self->{file} = $self->{dbdir} . "/gitcvs.$module.sqlite";
2094
2095     $self->{git_path} = $config . "/";
2096
2097     $self->{log} = $log;
2098
2099     die "Git repo '$self->{git_path}' doesn't exist" unless ( -d $self->{git_path} );
2100
2101     $self->{dbh} = DBI->connect("dbi:SQLite:dbname=" . $self->{file},"","");
2102
2103     $self->{tables} = {};
2104     foreach my $table ( $self->{dbh}->tables )
2105     {
2106         $table =~ s/^"//;
2107         $table =~ s/"$//;
2108         $self->{tables}{$table} = 1;
2109     }
2110
2111     # Construct the revision table if required
2112     unless ( $self->{tables}{revision} )
2113     {
2114         $self->{dbh}->do("
2115             CREATE TABLE revision (
2116                 name       TEXT NOT NULL,
2117                 revision   INTEGER NOT NULL,
2118                 filehash   TEXT NOT NULL,
2119                 commithash TEXT NOT NULL,
2120                 author     TEXT NOT NULL,
2121                 modified   TEXT NOT NULL,
2122                 mode       TEXT NOT NULL
2123             )
2124         ");
2125         $self->{dbh}->do("
2126             CREATE INDEX revision_ix1
2127             ON revision (name,revision)
2128         ");
2129         $self->{dbh}->do("
2130             CREATE INDEX revision_ix2
2131             ON revision (name,commithash)
2132         ");
2133     }
2134
2135     # Construct the head table if required
2136     unless ( $self->{tables}{head} )
2137     {
2138         $self->{dbh}->do("
2139             CREATE TABLE head (
2140                 name       TEXT NOT NULL,
2141                 revision   INTEGER NOT NULL,
2142                 filehash   TEXT NOT NULL,
2143                 commithash TEXT NOT NULL,
2144                 author     TEXT NOT NULL,
2145                 modified   TEXT NOT NULL,
2146                 mode       TEXT NOT NULL
2147             )
2148         ");
2149         $self->{dbh}->do("
2150             CREATE INDEX head_ix1
2151             ON head (name)
2152         ");
2153     }
2154
2155     # Construct the properties table if required
2156     unless ( $self->{tables}{properties} )
2157     {
2158         $self->{dbh}->do("
2159             CREATE TABLE properties (
2160                 key        TEXT NOT NULL PRIMARY KEY,
2161                 value      TEXT
2162             )
2163         ");
2164     }
2165
2166     # Construct the commitmsgs table if required
2167     unless ( $self->{tables}{commitmsgs} )
2168     {
2169         $self->{dbh}->do("
2170             CREATE TABLE commitmsgs (
2171                 key        TEXT NOT NULL PRIMARY KEY,
2172                 value      TEXT
2173             )
2174         ");
2175     }
2176
2177     return $self;
2178 }
2179
2180 =head2 update
2181
2182 =cut
2183 sub update
2184 {
2185     my $self = shift;
2186
2187     # first lets get the commit list
2188     $ENV{GIT_DIR} = $self->{git_path};
2189
2190     my $commitsha1 = `git rev-parse $self->{module}`;
2191     chomp $commitsha1;
2192
2193     my $commitinfo = `git cat-file commit $self->{module} 2>&1`;
2194     unless ( $commitinfo =~ /tree\s+[a-zA-Z0-9]{40}/ )
2195     {
2196         die("Invalid module '$self->{module}'");
2197     }
2198
2199
2200     my $git_log;
2201     my $lastcommit = $self->_get_prop("last_commit");
2202
2203     if (defined $lastcommit && $lastcommit eq $commitsha1) { # up-to-date
2204          return 1;
2205     }
2206
2207     # Start exclusive lock here...
2208     $self->{dbh}->begin_work() or die "Cannot lock database for BEGIN";
2209
2210     # TODO: log processing is memory bound
2211     # if we can parse into a 2nd file that is in reverse order
2212     # we can probably do something really efficient
2213     my @git_log_params = ('--pretty', '--parents', '--topo-order');
2214
2215     if (defined $lastcommit) {
2216         push @git_log_params, "$lastcommit..$self->{module}";
2217     } else {
2218         push @git_log_params, $self->{module};
2219     }
2220     # git-rev-list is the backend / plumbing version of git-log
2221     open(GITLOG, '-|', 'git-rev-list', @git_log_params) or die "Cannot call git-rev-list: $!";
2222
2223     my @commits;
2224
2225     my %commit = ();
2226
2227     while ( <GITLOG> )
2228     {
2229         chomp;
2230         if (m/^commit\s+(.*)$/) {
2231             # on ^commit lines put the just seen commit in the stack
2232             # and prime things for the next one
2233             if (keys %commit) {
2234                 my %copy = %commit;
2235                 unshift @commits, \%copy;
2236                 %commit = ();
2237             }
2238             my @parents = split(m/\s+/, $1);
2239             $commit{hash} = shift @parents;
2240             $commit{parents} = \@parents;
2241         } elsif (m/^(\w+?):\s+(.*)$/ && !exists($commit{message})) {
2242             # on rfc822-like lines seen before we see any message,
2243             # lowercase the entry and put it in the hash as key-value
2244             $commit{lc($1)} = $2;
2245         } else {
2246             # message lines - skip initial empty line
2247             # and trim whitespace
2248             if (!exists($commit{message}) && m/^\s*$/) {
2249                 # define it to mark the end of headers
2250                 $commit{message} = '';
2251                 next;
2252             }
2253             s/^\s+//; s/\s+$//; # trim ws
2254             $commit{message} .= $_ . "\n";
2255         }
2256     }
2257     close GITLOG;
2258
2259     unshift @commits, \%commit if ( keys %commit );
2260
2261     # Now all the commits are in the @commits bucket
2262     # ordered by time DESC. for each commit that needs processing,
2263     # determine whether it's following the last head we've seen or if
2264     # it's on its own branch, grab a file list, and add whatever's changed
2265     # NOTE: $lastcommit refers to the last commit from previous run
2266     #       $lastpicked is the last commit we picked in this run
2267     my $lastpicked;
2268     my $head = {};
2269     if (defined $lastcommit) {
2270         $lastpicked = $lastcommit;
2271     }
2272
2273     my $committotal = scalar(@commits);
2274     my $commitcount = 0;
2275
2276     # Load the head table into $head (for cached lookups during the update process)
2277     foreach my $file ( @{$self->gethead()} )
2278     {
2279         $head->{$file->{name}} = $file;
2280     }
2281
2282     foreach my $commit ( @commits )
2283     {
2284         $self->{log}->debug("GITCVS::updater - Processing commit $commit->{hash} (" . (++$commitcount) . " of $committotal)");
2285         if (defined $lastpicked)
2286         {
2287             if (!in_array($lastpicked, @{$commit->{parents}}))
2288             {
2289                 # skip, we'll see this delta
2290                 # as part of a merge later
2291                 # warn "skipping off-track  $commit->{hash}\n";
2292                 next;
2293             } elsif (@{$commit->{parents}} > 1) {
2294                 # it is a merge commit, for each parent that is
2295                 # not $lastpicked, see if we can get a log
2296                 # from the merge-base to that parent to put it
2297                 # in the message as a merge summary.
2298                 my @parents = @{$commit->{parents}};
2299                 foreach my $parent (@parents) {
2300                     # git-merge-base can potentially (but rarely) throw
2301                     # several candidate merge bases. let's assume
2302                     # that the first one is the best one.
2303                     if ($parent eq $lastpicked) {
2304                         next;
2305                     }
2306                     open my $p, 'git-merge-base '. $lastpicked . ' '
2307                     . $parent . '|';
2308                     my @output = (<$p>);
2309                     close $p;
2310                     my $base = join('', @output);
2311                     chomp $base;
2312                     if ($base) {
2313                         my @merged;
2314                         # print "want to log between  $base $parent \n";
2315                         open(GITLOG, '-|', 'git-log', "$base..$parent")
2316                         or die "Cannot call git-log: $!";
2317                         my $mergedhash;
2318                         while (<GITLOG>) {
2319                             chomp;
2320                             if (!defined $mergedhash) {
2321                                 if (m/^commit\s+(.+)$/) {
2322                                     $mergedhash = $1;
2323                                 } else {
2324                                     next;
2325                                 }
2326                             } else {
2327                                 # grab the first line that looks non-rfc822
2328                                 # aka has content after leading space
2329                                 if (m/^\s+(\S.*)$/) {
2330                                     my $title = $1;
2331                                     $title = substr($title,0,100); # truncate
2332                                     unshift @merged, "$mergedhash $title";
2333                                     undef $mergedhash;
2334                                 }
2335                             }
2336                         }
2337                         close GITLOG;
2338                         if (@merged) {
2339                             $commit->{mergemsg} = $commit->{message};
2340                             $commit->{mergemsg} .= "\nSummary of merged commits:\n\n";
2341                             foreach my $summary (@merged) {
2342                                 $commit->{mergemsg} .= "\t$summary\n";
2343                             }
2344                             $commit->{mergemsg} .= "\n\n";
2345                             # print "Message for $commit->{hash} \n$commit->{mergemsg}";
2346                         }
2347                     }
2348                 }
2349             }
2350         }
2351
2352         # convert the date to CVS-happy format
2353         $commit->{date} = "$2 $1 $4 $3 $5" if ( $commit->{date} =~ /^\w+\s+(\w+)\s+(\d+)\s+(\d+:\d+:\d+)\s+(\d+)\s+([+-]\d+)$/ );
2354
2355         if ( defined ( $lastpicked ) )
2356         {
2357             my $filepipe = open(FILELIST, '-|', 'git-diff-tree', '-z', '-r', $lastpicked, $commit->{hash}) or die("Cannot call git-diff-tree : $!");
2358             local ($/) = "\0";
2359             while ( <FILELIST> )
2360             {
2361                 chomp;
2362                 unless ( /^:\d{6}\s+\d{3}(\d)\d{2}\s+[a-zA-Z0-9]{40}\s+([a-zA-Z0-9]{40})\s+(\w)$/o )
2363                 {
2364                     die("Couldn't process git-diff-tree line : $_");
2365                 }
2366                 my ($mode, $hash, $change) = ($1, $2, $3);
2367                 my $name = <FILELIST>;
2368                 chomp($name);
2369
2370                 # $log->debug("File mode=$mode, hash=$hash, change=$change, name=$name");
2371
2372                 my $git_perms = "";
2373                 $git_perms .= "r" if ( $mode & 4 );
2374                 $git_perms .= "w" if ( $mode & 2 );
2375                 $git_perms .= "x" if ( $mode & 1 );
2376                 $git_perms = "rw" if ( $git_perms eq "" );
2377
2378                 if ( $change eq "D" )
2379                 {
2380                     #$log->debug("DELETE   $name");
2381                     $head->{$name} = {
2382                         name => $name,
2383                         revision => $head->{$name}{revision} + 1,
2384                         filehash => "deleted",
2385                         commithash => $commit->{hash},
2386                         modified => $commit->{date},
2387                         author => $commit->{author},
2388                         mode => $git_perms,
2389                     };
2390                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2391                 }
2392                 elsif ( $change eq "M" )
2393                 {
2394                     #$log->debug("MODIFIED $name");
2395                     $head->{$name} = {
2396                         name => $name,
2397                         revision => $head->{$name}{revision} + 1,
2398                         filehash => $hash,
2399                         commithash => $commit->{hash},
2400                         modified => $commit->{date},
2401                         author => $commit->{author},
2402                         mode => $git_perms,
2403                     };
2404                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2405                 }
2406                 elsif ( $change eq "A" )
2407                 {
2408                     #$log->debug("ADDED    $name");
2409                     $head->{$name} = {
2410                         name => $name,
2411                         revision => 1,
2412                         filehash => $hash,
2413                         commithash => $commit->{hash},
2414                         modified => $commit->{date},
2415                         author => $commit->{author},
2416                         mode => $git_perms,
2417                     };
2418                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2419                 }
2420                 else
2421                 {
2422                     $log->warn("UNKNOWN FILE CHANGE mode=$mode, hash=$hash, change=$change, name=$name");
2423                     die;
2424                 }
2425             }
2426             close FILELIST;
2427         } else {
2428             # this is used to detect files removed from the repo
2429             my $seen_files = {};
2430
2431             my $filepipe = open(FILELIST, '-|', 'git-ls-tree', '-z', '-r', $commit->{hash}) or die("Cannot call git-ls-tree : $!");
2432             local $/ = "\0";
2433             while ( <FILELIST> )
2434             {
2435                 chomp;
2436                 unless ( /^(\d+)\s+(\w+)\s+([a-zA-Z0-9]+)\t(.*)$/o )
2437                 {
2438                     die("Couldn't process git-ls-tree line : $_");
2439                 }
2440
2441                 my ( $git_perms, $git_type, $git_hash, $git_filename ) = ( $1, $2, $3, $4 );
2442
2443                 $seen_files->{$git_filename} = 1;
2444
2445                 my ( $oldhash, $oldrevision, $oldmode ) = (
2446                     $head->{$git_filename}{filehash},
2447                     $head->{$git_filename}{revision},
2448                     $head->{$git_filename}{mode}
2449                 );
2450
2451                 if ( $git_perms =~ /^\d\d\d(\d)\d\d/o )
2452                 {
2453                     $git_perms = "";
2454                     $git_perms .= "r" if ( $1 & 4 );
2455                     $git_perms .= "w" if ( $1 & 2 );
2456                     $git_perms .= "x" if ( $1 & 1 );
2457                 } else {
2458                     $git_perms = "rw";
2459                 }
2460
2461                 # unless the file exists with the same hash, we need to update it ...
2462                 unless ( defined($oldhash) and $oldhash eq $git_hash and defined($oldmode) and $oldmode eq $git_perms )
2463                 {
2464                     my $newrevision = ( $oldrevision or 0 ) + 1;
2465
2466                     $head->{$git_filename} = {
2467                         name => $git_filename,
2468                         revision => $newrevision,
2469                         filehash => $git_hash,
2470                         commithash => $commit->{hash},
2471                         modified => $commit->{date},
2472                         author => $commit->{author},
2473                         mode => $git_perms,
2474                     };
2475
2476
2477                     $self->insert_rev($git_filename, $newrevision, $git_hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2478                 }
2479             }
2480             close FILELIST;
2481
2482             # Detect deleted files
2483             foreach my $file ( keys %$head )
2484             {
2485                 unless ( exists $seen_files->{$file} or $head->{$file}{filehash} eq "deleted" )
2486                 {
2487                     $head->{$file}{revision}++;
2488                     $head->{$file}{filehash} = "deleted";
2489                     $head->{$file}{commithash} = $commit->{hash};
2490                     $head->{$file}{modified} = $commit->{date};
2491                     $head->{$file}{author} = $commit->{author};
2492
2493                     $self->insert_rev($file, $head->{$file}{revision}, $head->{$file}{filehash}, $commit->{hash}, $commit->{date}, $commit->{author}, $head->{$file}{mode});
2494                 }
2495             }
2496             # END : "Detect deleted files"
2497         }
2498
2499
2500         if (exists $commit->{mergemsg})
2501         {
2502             $self->insert_mergelog($commit->{hash}, $commit->{mergemsg});
2503         }
2504
2505         $lastpicked = $commit->{hash};
2506
2507         $self->_set_prop("last_commit", $commit->{hash});
2508     }
2509
2510     $self->delete_head();
2511     foreach my $file ( keys %$head )
2512     {
2513         $self->insert_head(
2514             $file,
2515             $head->{$file}{revision},
2516             $head->{$file}{filehash},
2517             $head->{$file}{commithash},
2518             $head->{$file}{modified},
2519             $head->{$file}{author},
2520             $head->{$file}{mode},
2521         );
2522     }
2523     # invalidate the gethead cache
2524     $self->{gethead_cache} = undef;
2525
2526
2527     # Ending exclusive lock here
2528     $self->{dbh}->commit() or die "Failed to commit changes to SQLite";
2529 }
2530
2531 sub insert_rev
2532 {
2533     my $self = shift;
2534     my $name = shift;
2535     my $revision = shift;
2536     my $filehash = shift;
2537     my $commithash = shift;
2538     my $modified = shift;
2539     my $author = shift;
2540     my $mode = shift;
2541
2542     my $insert_rev = $self->{dbh}->prepare_cached("INSERT INTO revision (name, revision, filehash, commithash, modified, author, mode) VALUES (?,?,?,?,?,?,?)",{},1);
2543     $insert_rev->execute($name, $revision, $filehash, $commithash, $modified, $author, $mode);
2544 }
2545
2546 sub insert_mergelog
2547 {
2548     my $self = shift;
2549     my $key = shift;
2550     my $value = shift;
2551
2552     my $insert_mergelog = $self->{dbh}->prepare_cached("INSERT INTO commitmsgs (key, value) VALUES (?,?)",{},1);
2553     $insert_mergelog->execute($key, $value);
2554 }
2555
2556 sub delete_head
2557 {
2558     my $self = shift;
2559
2560     my $delete_head = $self->{dbh}->prepare_cached("DELETE FROM head",{},1);
2561     $delete_head->execute();
2562 }
2563
2564 sub insert_head
2565 {
2566     my $self = shift;
2567     my $name = shift;
2568     my $revision = shift;
2569     my $filehash = shift;
2570     my $commithash = shift;
2571     my $modified = shift;
2572     my $author = shift;
2573     my $mode = shift;
2574
2575     my $insert_head = $self->{dbh}->prepare_cached("INSERT INTO head (name, revision, filehash, commithash, modified, author, mode) VALUES (?,?,?,?,?,?,?)",{},1);
2576     $insert_head->execute($name, $revision, $filehash, $commithash, $modified, $author, $mode);
2577 }
2578
2579 sub _headrev
2580 {
2581     my $self = shift;
2582     my $filename = shift;
2583
2584     my $db_query = $self->{dbh}->prepare_cached("SELECT filehash, revision, mode FROM head WHERE name=?",{},1);
2585     $db_query->execute($filename);
2586     my ( $hash, $revision, $mode ) = $db_query->fetchrow_array;
2587
2588     return ( $hash, $revision, $mode );
2589 }
2590
2591 sub _get_prop
2592 {
2593     my $self = shift;
2594     my $key = shift;
2595
2596     my $db_query = $self->{dbh}->prepare_cached("SELECT value FROM properties WHERE key=?",{},1);
2597     $db_query->execute($key);
2598     my ( $value ) = $db_query->fetchrow_array;
2599
2600     return $value;
2601 }
2602
2603 sub _set_prop
2604 {
2605     my $self = shift;
2606     my $key = shift;
2607     my $value = shift;
2608
2609     my $db_query = $self->{dbh}->prepare_cached("UPDATE properties SET value=? WHERE key=?",{},1);
2610     $db_query->execute($value, $key);
2611
2612     unless ( $db_query->rows )
2613     {
2614         $db_query = $self->{dbh}->prepare_cached("INSERT INTO properties (key, value) VALUES (?,?)",{},1);
2615         $db_query->execute($key, $value);
2616     }
2617
2618     return $value;
2619 }
2620
2621 =head2 gethead
2622
2623 =cut
2624
2625 sub gethead
2626 {
2627     my $self = shift;
2628
2629     return $self->{gethead_cache} if ( defined ( $self->{gethead_cache} ) );
2630
2631     my $db_query = $self->{dbh}->prepare_cached("SELECT name, filehash, mode, revision, modified, commithash, author FROM head ORDER BY name ASC",{},1);
2632     $db_query->execute();
2633
2634     my $tree = [];
2635     while ( my $file = $db_query->fetchrow_hashref )
2636     {
2637         push @$tree, $file;
2638     }
2639
2640     $self->{gethead_cache} = $tree;
2641
2642     return $tree;
2643 }
2644
2645 =head2 getlog
2646
2647 =cut
2648
2649 sub getlog
2650 {
2651     my $self = shift;
2652     my $filename = shift;
2653
2654     my $db_query = $self->{dbh}->prepare_cached("SELECT name, filehash, author, mode, revision, modified, commithash FROM revision WHERE name=? ORDER BY revision DESC",{},1);
2655     $db_query->execute($filename);
2656
2657     my $tree = [];
2658     while ( my $file = $db_query->fetchrow_hashref )
2659     {
2660         push @$tree, $file;
2661     }
2662
2663     return $tree;
2664 }
2665
2666 =head2 getmeta
2667
2668 This function takes a filename (with path) argument and returns a hashref of
2669 metadata for that file.
2670
2671 =cut
2672
2673 sub getmeta
2674 {
2675     my $self = shift;
2676     my $filename = shift;
2677     my $revision = shift;
2678
2679     my $db_query;
2680     if ( defined($revision) and $revision =~ /^\d+$/ )
2681     {
2682         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM revision WHERE name=? AND revision=?",{},1);
2683         $db_query->execute($filename, $revision);
2684     }
2685     elsif ( defined($revision) and $revision =~ /^[a-zA-Z0-9]{40}$/ )
2686     {
2687         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM revision WHERE name=? AND commithash=?",{},1);
2688         $db_query->execute($filename, $revision);
2689     } else {
2690         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM head WHERE name=?",{},1);
2691         $db_query->execute($filename);
2692     }
2693
2694     return $db_query->fetchrow_hashref;
2695 }
2696
2697 =head2 commitmessage
2698
2699 this function takes a commithash and returns the commit message for that commit
2700
2701 =cut
2702 sub commitmessage
2703 {
2704     my $self = shift;
2705     my $commithash = shift;
2706
2707     die("Need commithash") unless ( defined($commithash) and $commithash =~ /^[a-zA-Z0-9]{40}$/ );
2708
2709     my $db_query;
2710     $db_query = $self->{dbh}->prepare_cached("SELECT value FROM commitmsgs WHERE key=?",{},1);
2711     $db_query->execute($commithash);
2712
2713     my ( $message ) = $db_query->fetchrow_array;
2714
2715     if ( defined ( $message ) )
2716     {
2717         $message .= " " if ( $message =~ /\n$/ );
2718         return $message;
2719     }
2720
2721     my @lines = safe_pipe_capture("git-cat-file", "commit", $commithash);
2722     shift @lines while ( $lines[0] =~ /\S/ );
2723     $message = join("",@lines);
2724     $message .= " " if ( $message =~ /\n$/ );
2725     return $message;
2726 }
2727
2728 =head2 gethistory
2729
2730 This function takes a filename (with path) argument and returns an arrayofarrays
2731 containing revision,filehash,commithash ordered by revision descending
2732
2733 =cut
2734 sub gethistory
2735 {
2736     my $self = shift;
2737     my $filename = shift;
2738
2739     my $db_query;
2740     $db_query = $self->{dbh}->prepare_cached("SELECT revision, filehash, commithash FROM revision WHERE name=? ORDER BY revision DESC",{},1);
2741     $db_query->execute($filename);
2742
2743     return $db_query->fetchall_arrayref;
2744 }
2745
2746 =head2 gethistorydense
2747
2748 This function takes a filename (with path) argument and returns an arrayofarrays
2749 containing revision,filehash,commithash ordered by revision descending.
2750
2751 This version of gethistory skips deleted entries -- so it is useful for annotate.
2752 The 'dense' part is a reference to a '--dense' option available for git-rev-list
2753 and other git tools that depend on it.
2754
2755 =cut
2756 sub gethistorydense
2757 {
2758     my $self = shift;
2759     my $filename = shift;
2760
2761     my $db_query;
2762     $db_query = $self->{dbh}->prepare_cached("SELECT revision, filehash, commithash FROM revision WHERE name=? AND filehash!='deleted' ORDER BY revision DESC",{},1);
2763     $db_query->execute($filename);
2764
2765     return $db_query->fetchall_arrayref;
2766 }
2767
2768 =head2 in_array()
2769
2770 from Array::PAT - mimics the in_array() function
2771 found in PHP. Yuck but works for small arrays.
2772
2773 =cut
2774 sub in_array
2775 {
2776     my ($check, @array) = @_;
2777     my $retval = 0;
2778     foreach my $test (@array){
2779         if($check eq $test){
2780             $retval =  1;
2781         }
2782     }
2783     return $retval;
2784 }
2785
2786 =head2 safe_pipe_capture
2787
2788 an alternative to `command` that allows input to be passed as an array
2789 to work around shell problems with weird characters in arguments
2790
2791 =cut
2792 sub safe_pipe_capture {
2793
2794     my @output;
2795
2796     if (my $pid = open my $child, '-|') {
2797         @output = (<$child>);
2798         close $child or die join(' ',@_).": $! $?";
2799     } else {
2800         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
2801     }
2802     return wantarray ? @output : join('',@output);
2803 }
2804
2805
2806 1;