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 $ENV{GIT_DIR} is not a valid repository\n";
 172        return 0;
 173     }
 174
 175     my @gitvars = `git-config -l`;
 176     if ($?) {
 177        print "E problems executing git-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-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     # Grab a handle to the SQLite db and do any necessary updates
1035     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1036     $updater->update();
1037
1038     my $tmpdir = tempdir ( DIR => $TEMP_DIR );
1039     my ( undef, $file_index ) = tempfile ( DIR => $TEMP_DIR, OPEN => 0 );
1040     $log->info("Lockless commit start, basing commit on '$tmpdir', index file is '$file_index'");
1041
1042     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
1043     $ENV{GIT_INDEX_FILE} = $file_index;
1044
1045     # Remember where the head was at the beginning.
1046     my $parenthash = `git show-ref -s refs/heads/$state->{module}`;
1047     chomp $parenthash;
1048     if ($parenthash !~ /^[0-9a-f]{40}$/) {
1049             print "error 1 pserver cannot find the current HEAD of module";
1050             exit;
1051     }
1052
1053     chdir $tmpdir;
1054
1055     # populate the temporary index based
1056     system("git-read-tree", $parenthash);
1057     unless ($? == 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     my @committedfiles = ();
1064
1065     # foreach file specified on the command line ...
1066     foreach my $filename ( @{$state->{args}} )
1067     {
1068         my $committedfile = $filename;
1069         $filename = filecleanup($filename);
1070
1071         next unless ( exists $state->{entries}{$filename}{modified_filename} or not $state->{entries}{$filename}{unchanged} );
1072
1073         my $meta = $updater->getmeta($filename);
1074
1075         my $wrev = revparse($filename);
1076
1077         my ( $filepart, $dirpart ) = filenamesplit($filename);
1078
1079         # do a checkout of the file if it part of this tree
1080         if ($wrev) {
1081             system('git-checkout-index', '-f', '-u', $filename);
1082             unless ($? == 0) {
1083                 die "Error running git-checkout-index -f -u $filename : $!";
1084             }
1085         }
1086
1087         my $addflag = 0;
1088         my $rmflag = 0;
1089         $rmflag = 1 if ( defined($wrev) and $wrev < 0 );
1090         $addflag = 1 unless ( -e $filename );
1091
1092         # Do up to date checking
1093         unless ( $addflag or $wrev == $meta->{revision} or ( $rmflag and -$wrev == $meta->{revision} ) )
1094         {
1095             # fail everything if an up to date check fails
1096             print "error 1 Up to date check failed for $filename\n";
1097             chdir "/";
1098             exit;
1099         }
1100
1101         push @committedfiles, $committedfile;
1102         $log->info("Committing $filename");
1103
1104         system("mkdir","-p",$dirpart) unless ( -d $dirpart );
1105
1106         unless ( $rmflag )
1107         {
1108             $log->debug("rename $state->{entries}{$filename}{modified_filename} $filename");
1109             rename $state->{entries}{$filename}{modified_filename},$filename;
1110
1111             # Calculate modes to remove
1112             my $invmode = "";
1113             foreach ( qw (r w x) ) { $invmode .= $_ unless ( $state->{entries}{$filename}{modified_mode} =~ /$_/ ); }
1114
1115             $log->debug("chmod u+" . $state->{entries}{$filename}{modified_mode} . "-" . $invmode . " $filename");
1116             system("chmod","u+" .  $state->{entries}{$filename}{modified_mode} . "-" . $invmode, $filename);
1117         }
1118
1119         if ( $rmflag )
1120         {
1121             $log->info("Removing file '$filename'");
1122             unlink($filename);
1123             system("git-update-index", "--remove", $filename);
1124         }
1125         elsif ( $addflag )
1126         {
1127             $log->info("Adding file '$filename'");
1128             system("git-update-index", "--add", $filename);
1129         } else {
1130             $log->info("Updating file '$filename'");
1131             system("git-update-index", $filename);
1132         }
1133     }
1134
1135     unless ( scalar(@committedfiles) > 0 )
1136     {
1137         print "E No files to commit\n";
1138         print "ok\n";
1139         chdir "/";
1140         return;
1141     }
1142
1143     my $treehash = `git-write-tree`;
1144     chomp $treehash;
1145
1146     $log->debug("Treehash : $treehash, Parenthash : $parenthash");
1147
1148     # write our commit message out if we have one ...
1149     my ( $msg_fh, $msg_filename ) = tempfile( DIR => $TEMP_DIR );
1150     print $msg_fh $state->{opt}{m};# if ( exists ( $state->{opt}{m} ) );
1151     print $msg_fh "\n\nvia git-CVS emulator\n";
1152     close $msg_fh;
1153
1154     my $commithash = `git-commit-tree $treehash -p $parenthash < $msg_filename`;
1155     $log->info("Commit hash : $commithash");
1156
1157     unless ( $commithash =~ /[a-zA-Z0-9]{40}/ )
1158     {
1159         $log->warn("Commit failed (Invalid commit hash)");
1160         print "error 1 Commit failed (unknown reason)\n";
1161         chdir "/";
1162         exit;
1163     }
1164
1165         # Check that this is allowed, just as we would with a receive-pack
1166         my @cmd = ( $ENV{GIT_DIR}.'hooks/update', "refs/heads/$state->{module}",
1167                         $parenthash, $commithash );
1168         if( -x $cmd[0] ) {
1169                 unless( system( @cmd ) == 0 )
1170                 {
1171                         $log->warn("Commit failed (update hook declined to update ref)");
1172                         print "error 1 Commit failed (update hook declined)\n";
1173                         chdir "/";
1174                         exit;
1175                 }
1176         }
1177
1178         if (system(qw(git update-ref -m), "cvsserver ci",
1179                         "refs/heads/$state->{module}", $commithash, $parenthash)) {
1180                 $log->warn("update-ref for $state->{module} failed.");
1181                 print "error 1 Cannot commit -- update first\n";
1182                 exit;
1183         }
1184
1185     $updater->update();
1186
1187     # foreach file specified on the command line ...
1188     foreach my $filename ( @committedfiles )
1189     {
1190         $filename = filecleanup($filename);
1191
1192         my $meta = $updater->getmeta($filename);
1193         unless (defined $meta->{revision}) {
1194           $meta->{revision} = 1;
1195         }
1196
1197         my ( $filepart, $dirpart ) = filenamesplit($filename, 1);
1198
1199         $log->debug("Checked-in $dirpart : $filename");
1200
1201         if ( defined $meta->{filehash} && $meta->{filehash} eq "deleted" )
1202         {
1203             print "Remove-entry $dirpart\n";
1204             print "$filename\n";
1205         } else {
1206             print "Checked-in $dirpart\n";
1207             print "$filename\n";
1208             print "/$filepart/1.$meta->{revision}///\n";
1209         }
1210     }
1211
1212     chdir "/";
1213     print "ok\n";
1214 }
1215
1216 sub req_status
1217 {
1218     my ( $cmd, $data ) = @_;
1219
1220     argsplit("status");
1221
1222     $log->info("req_status : " . ( defined($data) ? $data : "[NULL]" ));
1223     #$log->debug("status state : " . Dumper($state));
1224
1225     # Grab a handle to the SQLite db and do any necessary updates
1226     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1227     $updater->update();
1228
1229     # if no files were specified, we need to work out what files we should be providing status on ...
1230     argsfromdir($updater);
1231
1232     # foreach file specified on the command line ...
1233     foreach my $filename ( @{$state->{args}} )
1234     {
1235         $filename = filecleanup($filename);
1236
1237         my $meta = $updater->getmeta($filename);
1238         my $oldmeta = $meta;
1239
1240         my $wrev = revparse($filename);
1241
1242         # If the working copy is an old revision, lets get that version too for comparison.
1243         if ( defined($wrev) and $wrev != $meta->{revision} )
1244         {
1245             $oldmeta = $updater->getmeta($filename, $wrev);
1246         }
1247
1248         # TODO : All possible statuses aren't yet implemented
1249         my $status;
1250         # Files are up to date if the working copy and repo copy have the same revision, and the working copy is unmodified
1251         $status = "Up-to-date" if ( defined ( $wrev ) and defined($meta->{revision}) and $wrev == $meta->{revision}
1252                                     and
1253                                     ( ( $state->{entries}{$filename}{unchanged} and ( not defined ( $state->{entries}{$filename}{conflict} ) or $state->{entries}{$filename}{conflict} !~ /^\+=/ ) )
1254                                       or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $meta->{filehash} ) )
1255                                    );
1256
1257         # Need checkout if the working copy has an older revision than the repo copy, and the working copy is unmodified
1258         $status ||= "Needs Checkout" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and $meta->{revision} > $wrev
1259                                           and
1260                                           ( $state->{entries}{$filename}{unchanged}
1261                                             or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $oldmeta->{filehash} ) )
1262                                         );
1263
1264         # Need checkout if it exists in the repo but doesn't have a working copy
1265         $status ||= "Needs Checkout" if ( not defined ( $wrev ) and defined ( $meta->{revision} ) );
1266
1267         # Locally modified if working copy and repo copy have the same revision but there are local changes
1268         $status ||= "Locally Modified" if ( defined ( $wrev ) and defined($meta->{revision}) and $wrev == $meta->{revision} and $state->{entries}{$filename}{modified_filename} );
1269
1270         # Needs Merge if working copy revision is less than repo copy and there are local changes
1271         $status ||= "Needs Merge" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and $meta->{revision} > $wrev and $state->{entries}{$filename}{modified_filename} );
1272
1273         $status ||= "Locally Added" if ( defined ( $state->{entries}{$filename}{revision} ) and not defined ( $meta->{revision} ) );
1274         $status ||= "Locally Removed" if ( defined ( $wrev ) and defined ( $meta->{revision} ) and -$wrev == $meta->{revision} );
1275         $status ||= "Unresolved Conflict" if ( defined ( $state->{entries}{$filename}{conflict} ) and $state->{entries}{$filename}{conflict} =~ /^\+=/ );
1276         $status ||= "File had conflicts on merge" if ( 0 );
1277
1278         $status ||= "Unknown";
1279
1280         print "M ===================================================================\n";
1281         print "M File: $filename\tStatus: $status\n";
1282         if ( defined($state->{entries}{$filename}{revision}) )
1283         {
1284             print "M Working revision:\t" . $state->{entries}{$filename}{revision} . "\n";
1285         } else {
1286             print "M Working revision:\tNo entry for $filename\n";
1287         }
1288         if ( defined($meta->{revision}) )
1289         {
1290             print "M Repository revision:\t1." . $meta->{revision} . "\t$state->{repository}/$filename,v\n";
1291             print "M Sticky Tag:\t\t(none)\n";
1292             print "M Sticky Date:\t\t(none)\n";
1293             print "M Sticky Options:\t\t(none)\n";
1294         } else {
1295             print "M Repository revision:\tNo revision control file\n";
1296         }
1297         print "M\n";
1298     }
1299
1300     print "ok\n";
1301 }
1302
1303 sub req_diff
1304 {
1305     my ( $cmd, $data ) = @_;
1306
1307     argsplit("diff");
1308
1309     $log->debug("req_diff : " . ( defined($data) ? $data : "[NULL]" ));
1310     #$log->debug("status state : " . Dumper($state));
1311
1312     my ($revision1, $revision2);
1313     if ( defined ( $state->{opt}{r} ) and ref $state->{opt}{r} eq "ARRAY" )
1314     {
1315         $revision1 = $state->{opt}{r}[0];
1316         $revision2 = $state->{opt}{r}[1];
1317     } else {
1318         $revision1 = $state->{opt}{r};
1319     }
1320
1321     $revision1 =~ s/^1\.// if ( defined ( $revision1 ) );
1322     $revision2 =~ s/^1\.// if ( defined ( $revision2 ) );
1323
1324     $log->debug("Diffing revisions " . ( defined($revision1) ? $revision1 : "[NULL]" ) . " and " . ( defined($revision2) ? $revision2 : "[NULL]" ) );
1325
1326     # Grab a handle to the SQLite db and do any necessary updates
1327     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1328     $updater->update();
1329
1330     # if no files were specified, we need to work out what files we should be providing status on ...
1331     argsfromdir($updater);
1332
1333     # foreach file specified on the command line ...
1334     foreach my $filename ( @{$state->{args}} )
1335     {
1336         $filename = filecleanup($filename);
1337
1338         my ( $fh, $file1, $file2, $meta1, $meta2, $filediff );
1339
1340         my $wrev = revparse($filename);
1341
1342         # We need _something_ to diff against
1343         next unless ( defined ( $wrev ) );
1344
1345         # if we have a -r switch, use it
1346         if ( defined ( $revision1 ) )
1347         {
1348             ( undef, $file1 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1349             $meta1 = $updater->getmeta($filename, $revision1);
1350             unless ( defined ( $meta1 ) and $meta1->{filehash} ne "deleted" )
1351             {
1352                 print "E File $filename at revision 1.$revision1 doesn't exist\n";
1353                 next;
1354             }
1355             transmitfile($meta1->{filehash}, $file1);
1356         }
1357         # otherwise we just use the working copy revision
1358         else
1359         {
1360             ( undef, $file1 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1361             $meta1 = $updater->getmeta($filename, $wrev);
1362             transmitfile($meta1->{filehash}, $file1);
1363         }
1364
1365         # if we have a second -r switch, use it too
1366         if ( defined ( $revision2 ) )
1367         {
1368             ( undef, $file2 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1369             $meta2 = $updater->getmeta($filename, $revision2);
1370
1371             unless ( defined ( $meta2 ) and $meta2->{filehash} ne "deleted" )
1372             {
1373                 print "E File $filename at revision 1.$revision2 doesn't exist\n";
1374                 next;
1375             }
1376
1377             transmitfile($meta2->{filehash}, $file2);
1378         }
1379         # otherwise we just use the working copy
1380         else
1381         {
1382             $file2 = $state->{entries}{$filename}{modified_filename};
1383         }
1384
1385         # if we have been given -r, and we don't have a $file2 yet, lets get one
1386         if ( defined ( $revision1 ) and not defined ( $file2 ) )
1387         {
1388             ( undef, $file2 ) = tempfile( DIR => $TEMP_DIR, OPEN => 0 );
1389             $meta2 = $updater->getmeta($filename, $wrev);
1390             transmitfile($meta2->{filehash}, $file2);
1391         }
1392
1393         # We need to have retrieved something useful
1394         next unless ( defined ( $meta1 ) );
1395
1396         # Files to date if the working copy and repo copy have the same revision, and the working copy is unmodified
1397         next if ( not defined ( $meta2 ) and $wrev == $meta1->{revision}
1398                   and
1399                    ( ( $state->{entries}{$filename}{unchanged} and ( not defined ( $state->{entries}{$filename}{conflict} ) or $state->{entries}{$filename}{conflict} !~ /^\+=/ ) )
1400                      or ( defined($state->{entries}{$filename}{modified_hash}) and $state->{entries}{$filename}{modified_hash} eq $meta1->{filehash} ) )
1401                   );
1402
1403         # Apparently we only show diffs for locally modified files
1404         next unless ( defined($meta2) or defined ( $state->{entries}{$filename}{modified_filename} ) );
1405
1406         print "M Index: $filename\n";
1407         print "M ===================================================================\n";
1408         print "M RCS file: $state->{CVSROOT}/$state->{module}/$filename,v\n";
1409         print "M retrieving revision 1.$meta1->{revision}\n" if ( defined ( $meta1 ) );
1410         print "M retrieving revision 1.$meta2->{revision}\n" if ( defined ( $meta2 ) );
1411         print "M diff ";
1412         foreach my $opt ( keys %{$state->{opt}} )
1413         {
1414             if ( ref $state->{opt}{$opt} eq "ARRAY" )
1415             {
1416                 foreach my $value ( @{$state->{opt}{$opt}} )
1417                 {
1418                     print "-$opt $value ";
1419                 }
1420             } else {
1421                 print "-$opt ";
1422                 print "$state->{opt}{$opt} " if ( defined ( $state->{opt}{$opt} ) );
1423             }
1424         }
1425         print "$filename\n";
1426
1427         $log->info("Diffing $filename -r $meta1->{revision} -r " . ( $meta2->{revision} or "workingcopy" ));
1428
1429         ( $fh, $filediff ) = tempfile ( DIR => $TEMP_DIR );
1430
1431         if ( exists $state->{opt}{u} )
1432         {
1433             system("diff -u -L '$filename revision 1.$meta1->{revision}' -L '$filename " . ( defined($meta2->{revision}) ? "revision 1.$meta2->{revision}" : "working copy" ) . "' $file1 $file2 > $filediff");
1434         } else {
1435             system("diff $file1 $file2 > $filediff");
1436         }
1437
1438         while ( <$fh> )
1439         {
1440             print "M $_";
1441         }
1442         close $fh;
1443     }
1444
1445     print "ok\n";
1446 }
1447
1448 sub req_log
1449 {
1450     my ( $cmd, $data ) = @_;
1451
1452     argsplit("log");
1453
1454     $log->debug("req_log : " . ( defined($data) ? $data : "[NULL]" ));
1455     #$log->debug("log state : " . Dumper($state));
1456
1457     my ( $minrev, $maxrev );
1458     if ( defined ( $state->{opt}{r} ) and $state->{opt}{r} =~ /([\d.]+)?(::?)([\d.]+)?/ )
1459     {
1460         my $control = $2;
1461         $minrev = $1;
1462         $maxrev = $3;
1463         $minrev =~ s/^1\.// if ( defined ( $minrev ) );
1464         $maxrev =~ s/^1\.// if ( defined ( $maxrev ) );
1465         $minrev++ if ( defined($minrev) and $control eq "::" );
1466     }
1467
1468     # Grab a handle to the SQLite db and do any necessary updates
1469     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1470     $updater->update();
1471
1472     # if no files were specified, we need to work out what files we should be providing status on ...
1473     argsfromdir($updater);
1474
1475     # foreach file specified on the command line ...
1476     foreach my $filename ( @{$state->{args}} )
1477     {
1478         $filename = filecleanup($filename);
1479
1480         my $headmeta = $updater->getmeta($filename);
1481
1482         my $revisions = $updater->getlog($filename);
1483         my $totalrevisions = scalar(@$revisions);
1484
1485         if ( defined ( $minrev ) )
1486         {
1487             $log->debug("Removing revisions less than $minrev");
1488             while ( scalar(@$revisions) > 0 and $revisions->[-1]{revision} < $minrev )
1489             {
1490                 pop @$revisions;
1491             }
1492         }
1493         if ( defined ( $maxrev ) )
1494         {
1495             $log->debug("Removing revisions greater than $maxrev");
1496             while ( scalar(@$revisions) > 0 and $revisions->[0]{revision} > $maxrev )
1497             {
1498                 shift @$revisions;
1499             }
1500         }
1501
1502         next unless ( scalar(@$revisions) );
1503
1504         print "M \n";
1505         print "M RCS file: $state->{CVSROOT}/$state->{module}/$filename,v\n";
1506         print "M Working file: $filename\n";
1507         print "M head: 1.$headmeta->{revision}\n";
1508         print "M branch:\n";
1509         print "M locks: strict\n";
1510         print "M access list:\n";
1511         print "M symbolic names:\n";
1512         print "M keyword substitution: kv\n";
1513         print "M total revisions: $totalrevisions;\tselected revisions: " . scalar(@$revisions) . "\n";
1514         print "M description:\n";
1515
1516         foreach my $revision ( @$revisions )
1517         {
1518             print "M ----------------------------\n";
1519             print "M revision 1.$revision->{revision}\n";
1520             # reformat the date for log output
1521             $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}) );
1522             $revision->{author} =~ s/\s+.*//;
1523             $revision->{author} =~ s/^(.{8}).*/$1/;
1524             print "M date: $revision->{modified};  author: $revision->{author};  state: " . ( $revision->{filehash} eq "deleted" ? "dead" : "Exp" ) . ";  lines: +2 -3\n";
1525             my $commitmessage = $updater->commitmessage($revision->{commithash});
1526             $commitmessage =~ s/^/M /mg;
1527             print $commitmessage . "\n";
1528         }
1529         print "M =============================================================================\n";
1530     }
1531
1532     print "ok\n";
1533 }
1534
1535 sub req_annotate
1536 {
1537     my ( $cmd, $data ) = @_;
1538
1539     argsplit("annotate");
1540
1541     $log->info("req_annotate : " . ( defined($data) ? $data : "[NULL]" ));
1542     #$log->debug("status state : " . Dumper($state));
1543
1544     # Grab a handle to the SQLite db and do any necessary updates
1545     my $updater = GITCVS::updater->new($state->{CVSROOT}, $state->{module}, $log);
1546     $updater->update();
1547
1548     # if no files were specified, we need to work out what files we should be providing annotate on ...
1549     argsfromdir($updater);
1550
1551     # we'll need a temporary checkout dir
1552     my $tmpdir = tempdir ( DIR => $TEMP_DIR );
1553     my ( undef, $file_index ) = tempfile ( DIR => $TEMP_DIR, OPEN => 0 );
1554     $log->info("Temp checkoutdir creation successful, basing annotate session work on '$tmpdir', index file is '$file_index'");
1555
1556     $ENV{GIT_DIR} = $state->{CVSROOT} . "/";
1557     $ENV{GIT_INDEX_FILE} = $file_index;
1558
1559     chdir $tmpdir;
1560
1561     # foreach file specified on the command line ...
1562     foreach my $filename ( @{$state->{args}} )
1563     {
1564         $filename = filecleanup($filename);
1565
1566         my $meta = $updater->getmeta($filename);
1567
1568         next unless ( $meta->{revision} );
1569
1570         # get all the commits that this file was in
1571         # in dense format -- aka skip dead revisions
1572         my $revisions   = $updater->gethistorydense($filename);
1573         my $lastseenin  = $revisions->[0][2];
1574
1575         # populate the temporary index based on the latest commit were we saw
1576         # the file -- but do it cheaply without checking out any files
1577         # TODO: if we got a revision from the client, use that instead
1578         # to look up the commithash in sqlite (still good to default to
1579         # the current head as we do now)
1580         system("git-read-tree", $lastseenin);
1581         unless ($? == 0)
1582         {
1583             die "Error running git-read-tree $lastseenin $file_index $!";
1584         }
1585         $log->info("Created index '$file_index' with commit $lastseenin - exit status $?");
1586
1587         # do a checkout of the file
1588         system('git-checkout-index', '-f', '-u', $filename);
1589         unless ($? == 0) {
1590             die "Error running git-checkout-index -f -u $filename : $!";
1591         }
1592
1593         $log->info("Annotate $filename");
1594
1595         # Prepare a file with the commits from the linearized
1596         # history that annotate should know about. This prevents
1597         # git-jsannotate telling us about commits we are hiding
1598         # from the client.
1599
1600         open(ANNOTATEHINTS, ">$tmpdir/.annotate_hints") or die "Error opening > $tmpdir/.annotate_hints $!";
1601         for (my $i=0; $i < @$revisions; $i++)
1602         {
1603             print ANNOTATEHINTS $revisions->[$i][2];
1604             if ($i+1 < @$revisions) { # have we got a parent?
1605                 print ANNOTATEHINTS ' ' . $revisions->[$i+1][2];
1606             }
1607             print ANNOTATEHINTS "\n";
1608         }
1609
1610         print ANNOTATEHINTS "\n";
1611         close ANNOTATEHINTS;
1612
1613         my $annotatecmd = 'git-annotate';
1614         open(ANNOTATE, "-|", $annotatecmd, '-l', '-S', "$tmpdir/.annotate_hints", $filename)
1615             or die "Error invoking $annotatecmd -l -S $tmpdir/.annotate_hints $filename : $!";
1616         my $metadata = {};
1617         print "E Annotations for $filename\n";
1618         print "E ***************\n";
1619         while ( <ANNOTATE> )
1620         {
1621             if (m/^([a-zA-Z0-9]{40})\t\([^\)]*\)(.*)$/i)
1622             {
1623                 my $commithash = $1;
1624                 my $data = $2;
1625                 unless ( defined ( $metadata->{$commithash} ) )
1626                 {
1627                     $metadata->{$commithash} = $updater->getmeta($filename, $commithash);
1628                     $metadata->{$commithash}{author} =~ s/\s+.*//;
1629                     $metadata->{$commithash}{author} =~ s/^(.{8}).*/$1/;
1630                     $metadata->{$commithash}{modified} = sprintf("%02d-%s-%02d", $1, $2, $3) if ( $metadata->{$commithash}{modified} =~ /^(\d+)\s(\w+)\s\d\d(\d\d)/ );
1631                 }
1632                 printf("M 1.%-5d      (%-8s %10s): %s\n",
1633                     $metadata->{$commithash}{revision},
1634                     $metadata->{$commithash}{author},
1635                     $metadata->{$commithash}{modified},
1636                     $data
1637                 );
1638             } else {
1639                 $log->warn("Error in annotate output! LINE: $_");
1640                 print "E Annotate error \n";
1641                 next;
1642             }
1643         }
1644         close ANNOTATE;
1645     }
1646
1647     # done; get out of the tempdir
1648     chdir "/";
1649
1650     print "ok\n";
1651
1652 }
1653
1654 # This method takes the state->{arguments} array and produces two new arrays.
1655 # The first is $state->{args} which is everything before the '--' argument, and
1656 # the second is $state->{files} which is everything after it.
1657 sub argsplit
1658 {
1659     return unless( defined($state->{arguments}) and ref $state->{arguments} eq "ARRAY" );
1660
1661     my $type = shift;
1662
1663     $state->{args} = [];
1664     $state->{files} = [];
1665     $state->{opt} = {};
1666
1667     if ( defined($type) )
1668     {
1669         my $opt = {};
1670         $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" );
1671         $opt = { v => 0, l => 0, R => 0 } if ( $type eq "status" );
1672         $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" );
1673         $opt = { l => 0, R => 0, k => 1, D => 1, D => 1, r => 2 } if ( $type eq "diff" );
1674         $opt = { c => 0, R => 0, l => 0, f => 0, F => 1, m => 1, r => 1 } if ( $type eq "ci" );
1675         $opt = { k => 1, m => 1 } if ( $type eq "add" );
1676         $opt = { f => 0, l => 0, R => 0 } if ( $type eq "remove" );
1677         $opt = { l => 0, b => 0, h => 0, R => 0, t => 0, N => 0, S => 0, r => 1, d => 1, s => 1, w => 1 } if ( $type eq "log" );
1678
1679
1680         while ( scalar ( @{$state->{arguments}} ) > 0 )
1681         {
1682             my $arg = shift @{$state->{arguments}};
1683
1684             next if ( $arg eq "--" );
1685             next unless ( $arg =~ /\S/ );
1686
1687             # if the argument looks like a switch
1688             if ( $arg =~ /^-(\w)(.*)/ )
1689             {
1690                 # if it's a switch that takes an argument
1691                 if ( $opt->{$1} )
1692                 {
1693                     # If this switch has already been provided
1694                     if ( $opt->{$1} > 1 and exists ( $state->{opt}{$1} ) )
1695                     {
1696                         $state->{opt}{$1} = [ $state->{opt}{$1} ];
1697                         if ( length($2) > 0 )
1698                         {
1699                             push @{$state->{opt}{$1}},$2;
1700                         } else {
1701                             push @{$state->{opt}{$1}}, shift @{$state->{arguments}};
1702                         }
1703                     } else {
1704                         # if there's extra data in the arg, use that as the argument for the switch
1705                         if ( length($2) > 0 )
1706                         {
1707                             $state->{opt}{$1} = $2;
1708                         } else {
1709                             $state->{opt}{$1} = shift @{$state->{arguments}};
1710                         }
1711                     }
1712                 } else {
1713                     $state->{opt}{$1} = undef;
1714                 }
1715             }
1716             else
1717             {
1718                 push @{$state->{args}}, $arg;
1719             }
1720         }
1721     }
1722     else
1723     {
1724         my $mode = 0;
1725
1726         foreach my $value ( @{$state->{arguments}} )
1727         {
1728             if ( $value eq "--" )
1729             {
1730                 $mode++;
1731                 next;
1732             }
1733             push @{$state->{args}}, $value if ( $mode == 0 );
1734             push @{$state->{files}}, $value if ( $mode == 1 );
1735         }
1736     }
1737 }
1738
1739 # This method uses $state->{directory} to populate $state->{args} with a list of filenames
1740 sub argsfromdir
1741 {
1742     my $updater = shift;
1743
1744     $state->{args} = [] if ( scalar(@{$state->{args}}) == 1 and $state->{args}[0] eq "." );
1745
1746     return if ( scalar ( @{$state->{args}} ) > 1 );
1747
1748     my @gethead = @{$updater->gethead};
1749
1750     # push added files
1751     foreach my $file (keys %{$state->{entries}}) {
1752         if ( exists $state->{entries}{$file}{revision} &&
1753                 $state->{entries}{$file}{revision} == 0 )
1754         {
1755             push @gethead, { name => $file, filehash => 'added' };
1756         }
1757     }
1758
1759     if ( scalar(@{$state->{args}}) == 1 )
1760     {
1761         my $arg = $state->{args}[0];
1762         $arg .= $state->{prependdir} if ( defined ( $state->{prependdir} ) );
1763
1764         $log->info("Only one arg specified, checking for directory expansion on '$arg'");
1765
1766         foreach my $file ( @gethead )
1767         {
1768             next if ( $file->{filehash} eq "deleted" and not defined ( $state->{entries}{$file->{name}} ) );
1769             next unless ( $file->{name} =~ /^$arg\// or $file->{name} eq $arg  );
1770             push @{$state->{args}}, $file->{name};
1771         }
1772
1773         shift @{$state->{args}} if ( scalar(@{$state->{args}}) > 1 );
1774     } else {
1775         $log->info("Only one arg specified, populating file list automatically");
1776
1777         $state->{args} = [];
1778
1779         foreach my $file ( @gethead )
1780         {
1781             next if ( $file->{filehash} eq "deleted" and not defined ( $state->{entries}{$file->{name}} ) );
1782             next unless ( $file->{name} =~ s/^$state->{prependdir}// );
1783             push @{$state->{args}}, $file->{name};
1784         }
1785     }
1786 }
1787
1788 # This method cleans up the $state variable after a command that uses arguments has run
1789 sub statecleanup
1790 {
1791     $state->{files} = [];
1792     $state->{args} = [];
1793     $state->{arguments} = [];
1794     $state->{entries} = {};
1795 }
1796
1797 sub revparse
1798 {
1799     my $filename = shift;
1800
1801     return undef unless ( defined ( $state->{entries}{$filename}{revision} ) );
1802
1803     return $1 if ( $state->{entries}{$filename}{revision} =~ /^1\.(\d+)/ );
1804     return -$1 if ( $state->{entries}{$filename}{revision} =~ /^-1\.(\d+)/ );
1805
1806     return undef;
1807 }
1808
1809 # This method takes a file hash and does a CVS "file transfer" which transmits the
1810 # size of the file, and then the file contents.
1811 # If a second argument $targetfile is given, the file is instead written out to
1812 # a file by the name of $targetfile
1813 sub transmitfile
1814 {
1815     my $filehash = shift;
1816     my $targetfile = shift;
1817
1818     if ( defined ( $filehash ) and $filehash eq "deleted" )
1819     {
1820         $log->warn("filehash is 'deleted'");
1821         return;
1822     }
1823
1824     die "Need filehash" unless ( defined ( $filehash ) and $filehash =~ /^[a-zA-Z0-9]{40}$/ );
1825
1826     my $type = `git-cat-file -t $filehash`;
1827     chomp $type;
1828
1829     die ( "Invalid type '$type' (expected 'blob')" ) unless ( defined ( $type ) and $type eq "blob" );
1830
1831     my $size = `git-cat-file -s $filehash`;
1832     chomp $size;
1833
1834     $log->debug("transmitfile($filehash) size=$size, type=$type");
1835
1836     if ( open my $fh, '-|', "git-cat-file", "blob", $filehash )
1837     {
1838         if ( defined ( $targetfile ) )
1839         {
1840             open NEWFILE, ">", $targetfile or die("Couldn't open '$targetfile' for writing : $!");
1841             print NEWFILE $_ while ( <$fh> );
1842             close NEWFILE;
1843         } else {
1844             print "$size\n";
1845             print while ( <$fh> );
1846         }
1847         close $fh or die ("Couldn't close filehandle for transmitfile()");
1848     } else {
1849         die("Couldn't execute git-cat-file");
1850     }
1851 }
1852
1853 # This method takes a file name, and returns ( $dirpart, $filepart ) which
1854 # refers to the directory portion and the file portion of the filename
1855 # respectively
1856 sub filenamesplit
1857 {
1858     my $filename = shift;
1859     my $fixforlocaldir = shift;
1860
1861     my ( $filepart, $dirpart ) = ( $filename, "." );
1862     ( $filepart, $dirpart ) = ( $2, $1 ) if ( $filename =~ /(.*)\/(.*)/ );
1863     $dirpart .= "/";
1864
1865     if ( $fixforlocaldir )
1866     {
1867         $dirpart =~ s/^$state->{prependdir}//;
1868     }
1869
1870     return ( $filepart, $dirpart );
1871 }
1872
1873 sub filecleanup
1874 {
1875     my $filename = shift;
1876
1877     return undef unless(defined($filename));
1878     if ( $filename =~ /^\// )
1879     {
1880         print "E absolute filenames '$filename' not supported by server\n";
1881         return undef;
1882     }
1883
1884     $filename =~ s/^\.\///g;
1885     $filename = $state->{prependdir} . $filename;
1886     return $filename;
1887 }
1888
1889 package GITCVS::log;
1890
1891 ####
1892 #### Copyright The Open University UK - 2006.
1893 ####
1894 #### Authors: Martyn Smith    <martyn@catalyst.net.nz>
1895 ####          Martin Langhoff <martin@catalyst.net.nz>
1896 ####
1897 ####
1898
1899 use strict;
1900 use warnings;
1901
1902 =head1 NAME
1903
1904 GITCVS::log
1905
1906 =head1 DESCRIPTION
1907
1908 This module provides very crude logging with a similar interface to
1909 Log::Log4perl
1910
1911 =head1 METHODS
1912
1913 =cut
1914
1915 =head2 new
1916
1917 Creates a new log object, optionally you can specify a filename here to
1918 indicate the file to log to. If no log file is specified, you can specify one
1919 later with method setfile, or indicate you no longer want logging with method
1920 nofile.
1921
1922 Until one of these methods is called, all log calls will buffer messages ready
1923 to write out.
1924
1925 =cut
1926 sub new
1927 {
1928     my $class = shift;
1929     my $filename = shift;
1930
1931     my $self = {};
1932
1933     bless $self, $class;
1934
1935     if ( defined ( $filename ) )
1936     {
1937         open $self->{fh}, ">>", $filename or die("Couldn't open '$filename' for writing : $!");
1938     }
1939
1940     return $self;
1941 }
1942
1943 =head2 setfile
1944
1945 This methods takes a filename, and attempts to open that file as the log file.
1946 If successful, all buffered data is written out to the file, and any further
1947 logging is written directly to the file.
1948
1949 =cut
1950 sub setfile
1951 {
1952     my $self = shift;
1953     my $filename = shift;
1954
1955     if ( defined ( $filename ) )
1956     {
1957         open $self->{fh}, ">>", $filename or die("Couldn't open '$filename' for writing : $!");
1958     }
1959
1960     return unless ( defined ( $self->{buffer} ) and ref $self->{buffer} eq "ARRAY" );
1961
1962     while ( my $line = shift @{$self->{buffer}} )
1963     {
1964         print {$self->{fh}} $line;
1965     }
1966 }
1967
1968 =head2 nofile
1969
1970 This method indicates no logging is going to be used. It flushes any entries in
1971 the internal buffer, and sets a flag to ensure no further data is put there.
1972
1973 =cut
1974 sub nofile
1975 {
1976     my $self = shift;
1977
1978     $self->{nolog} = 1;
1979
1980     return unless ( defined ( $self->{buffer} ) and ref $self->{buffer} eq "ARRAY" );
1981
1982     $self->{buffer} = [];
1983 }
1984
1985 =head2 _logopen
1986
1987 Internal method. Returns true if the log file is open, false otherwise.
1988
1989 =cut
1990 sub _logopen
1991 {
1992     my $self = shift;
1993
1994     return 1 if ( defined ( $self->{fh} ) and ref $self->{fh} eq "GLOB" );
1995     return 0;
1996 }
1997
1998 =head2 debug info warn fatal
1999
2000 These four methods are wrappers to _log. They provide the actual interface for
2001 logging data.
2002
2003 =cut
2004 sub debug { my $self = shift; $self->_log("debug", @_); }
2005 sub info  { my $self = shift; $self->_log("info" , @_); }
2006 sub warn  { my $self = shift; $self->_log("warn" , @_); }
2007 sub fatal { my $self = shift; $self->_log("fatal", @_); }
2008
2009 =head2 _log
2010
2011 This is an internal method called by the logging functions. It generates a
2012 timestamp and pushes the logged line either to file, or internal buffer.
2013
2014 =cut
2015 sub _log
2016 {
2017     my $self = shift;
2018     my $level = shift;
2019
2020     return if ( $self->{nolog} );
2021
2022     my @time = localtime;
2023     my $timestring = sprintf("%4d-%02d-%02d %02d:%02d:%02d : %-5s",
2024         $time[5] + 1900,
2025         $time[4] + 1,
2026         $time[3],
2027         $time[2],
2028         $time[1],
2029         $time[0],
2030         uc $level,
2031     );
2032
2033     if ( $self->_logopen )
2034     {
2035         print {$self->{fh}} $timestring . " - " . join(" ",@_) . "\n";
2036     } else {
2037         push @{$self->{buffer}}, $timestring . " - " . join(" ",@_) . "\n";
2038     }
2039 }
2040
2041 =head2 DESTROY
2042
2043 This method simply closes the file handle if one is open
2044
2045 =cut
2046 sub DESTROY
2047 {
2048     my $self = shift;
2049
2050     if ( $self->_logopen )
2051     {
2052         close $self->{fh};
2053     }
2054 }
2055
2056 package GITCVS::updater;
2057
2058 ####
2059 #### Copyright The Open University UK - 2006.
2060 ####
2061 #### Authors: Martyn Smith    <martyn@catalyst.net.nz>
2062 ####          Martin Langhoff <martin@catalyst.net.nz>
2063 ####
2064 ####
2065
2066 use strict;
2067 use warnings;
2068 use DBI;
2069
2070 =head1 METHODS
2071
2072 =cut
2073
2074 =head2 new
2075
2076 =cut
2077 sub new
2078 {
2079     my $class = shift;
2080     my $config = shift;
2081     my $module = shift;
2082     my $log = shift;
2083
2084     die "Need to specify a git repository" unless ( defined($config) and -d $config );
2085     die "Need to specify a module" unless ( defined($module) );
2086
2087     $class = ref($class) || $class;
2088
2089     my $self = {};
2090
2091     bless $self, $class;
2092
2093     $self->{dbdir} = $config . "/";
2094     die "Database dir '$self->{dbdir}' isn't a directory" unless ( defined($self->{dbdir}) and -d $self->{dbdir} );
2095
2096     $self->{module} = $module;
2097     $self->{file} = $self->{dbdir} . "/gitcvs.$module.sqlite";
2098
2099     $self->{git_path} = $config . "/";
2100
2101     $self->{log} = $log;
2102
2103     die "Git repo '$self->{git_path}' doesn't exist" unless ( -d $self->{git_path} );
2104
2105     $self->{dbh} = DBI->connect("dbi:SQLite:dbname=" . $self->{file},"","");
2106
2107     $self->{tables} = {};
2108     foreach my $table ( $self->{dbh}->tables )
2109     {
2110         $table =~ s/^"//;
2111         $table =~ s/"$//;
2112         $self->{tables}{$table} = 1;
2113     }
2114
2115     # Construct the revision table if required
2116     unless ( $self->{tables}{revision} )
2117     {
2118         $self->{dbh}->do("
2119             CREATE TABLE revision (
2120                 name       TEXT NOT NULL,
2121                 revision   INTEGER NOT NULL,
2122                 filehash   TEXT NOT NULL,
2123                 commithash TEXT NOT NULL,
2124                 author     TEXT NOT NULL,
2125                 modified   TEXT NOT NULL,
2126                 mode       TEXT NOT NULL
2127             )
2128         ");
2129         $self->{dbh}->do("
2130             CREATE INDEX revision_ix1
2131             ON revision (name,revision)
2132         ");
2133         $self->{dbh}->do("
2134             CREATE INDEX revision_ix2
2135             ON revision (name,commithash)
2136         ");
2137     }
2138
2139     # Construct the head table if required
2140     unless ( $self->{tables}{head} )
2141     {
2142         $self->{dbh}->do("
2143             CREATE TABLE head (
2144                 name       TEXT NOT NULL,
2145                 revision   INTEGER NOT NULL,
2146                 filehash   TEXT NOT NULL,
2147                 commithash TEXT NOT NULL,
2148                 author     TEXT NOT NULL,
2149                 modified   TEXT NOT NULL,
2150                 mode       TEXT NOT NULL
2151             )
2152         ");
2153         $self->{dbh}->do("
2154             CREATE INDEX head_ix1
2155             ON head (name)
2156         ");
2157     }
2158
2159     # Construct the properties table if required
2160     unless ( $self->{tables}{properties} )
2161     {
2162         $self->{dbh}->do("
2163             CREATE TABLE properties (
2164                 key        TEXT NOT NULL PRIMARY KEY,
2165                 value      TEXT
2166             )
2167         ");
2168     }
2169
2170     # Construct the commitmsgs table if required
2171     unless ( $self->{tables}{commitmsgs} )
2172     {
2173         $self->{dbh}->do("
2174             CREATE TABLE commitmsgs (
2175                 key        TEXT NOT NULL PRIMARY KEY,
2176                 value      TEXT
2177             )
2178         ");
2179     }
2180
2181     return $self;
2182 }
2183
2184 =head2 update
2185
2186 =cut
2187 sub update
2188 {
2189     my $self = shift;
2190
2191     # first lets get the commit list
2192     $ENV{GIT_DIR} = $self->{git_path};
2193
2194     my $commitsha1 = `git rev-parse $self->{module}`;
2195     chomp $commitsha1;
2196
2197     my $commitinfo = `git cat-file commit $self->{module} 2>&1`;
2198     unless ( $commitinfo =~ /tree\s+[a-zA-Z0-9]{40}/ )
2199     {
2200         die("Invalid module '$self->{module}'");
2201     }
2202
2203
2204     my $git_log;
2205     my $lastcommit = $self->_get_prop("last_commit");
2206
2207     if (defined $lastcommit && $lastcommit eq $commitsha1) { # up-to-date
2208          return 1;
2209     }
2210
2211     # Start exclusive lock here...
2212     $self->{dbh}->begin_work() or die "Cannot lock database for BEGIN";
2213
2214     # TODO: log processing is memory bound
2215     # if we can parse into a 2nd file that is in reverse order
2216     # we can probably do something really efficient
2217     my @git_log_params = ('--pretty', '--parents', '--topo-order');
2218
2219     if (defined $lastcommit) {
2220         push @git_log_params, "$lastcommit..$self->{module}";
2221     } else {
2222         push @git_log_params, $self->{module};
2223     }
2224     # git-rev-list is the backend / plumbing version of git-log
2225     open(GITLOG, '-|', 'git-rev-list', @git_log_params) or die "Cannot call git-rev-list: $!";
2226
2227     my @commits;
2228
2229     my %commit = ();
2230
2231     while ( <GITLOG> )
2232     {
2233         chomp;
2234         if (m/^commit\s+(.*)$/) {
2235             # on ^commit lines put the just seen commit in the stack
2236             # and prime things for the next one
2237             if (keys %commit) {
2238                 my %copy = %commit;
2239                 unshift @commits, \%copy;
2240                 %commit = ();
2241             }
2242             my @parents = split(m/\s+/, $1);
2243             $commit{hash} = shift @parents;
2244             $commit{parents} = \@parents;
2245         } elsif (m/^(\w+?):\s+(.*)$/ && !exists($commit{message})) {
2246             # on rfc822-like lines seen before we see any message,
2247             # lowercase the entry and put it in the hash as key-value
2248             $commit{lc($1)} = $2;
2249         } else {
2250             # message lines - skip initial empty line
2251             # and trim whitespace
2252             if (!exists($commit{message}) && m/^\s*$/) {
2253                 # define it to mark the end of headers
2254                 $commit{message} = '';
2255                 next;
2256             }
2257             s/^\s+//; s/\s+$//; # trim ws
2258             $commit{message} .= $_ . "\n";
2259         }
2260     }
2261     close GITLOG;
2262
2263     unshift @commits, \%commit if ( keys %commit );
2264
2265     # Now all the commits are in the @commits bucket
2266     # ordered by time DESC. for each commit that needs processing,
2267     # determine whether it's following the last head we've seen or if
2268     # it's on its own branch, grab a file list, and add whatever's changed
2269     # NOTE: $lastcommit refers to the last commit from previous run
2270     #       $lastpicked is the last commit we picked in this run
2271     my $lastpicked;
2272     my $head = {};
2273     if (defined $lastcommit) {
2274         $lastpicked = $lastcommit;
2275     }
2276
2277     my $committotal = scalar(@commits);
2278     my $commitcount = 0;
2279
2280     # Load the head table into $head (for cached lookups during the update process)
2281     foreach my $file ( @{$self->gethead()} )
2282     {
2283         $head->{$file->{name}} = $file;
2284     }
2285
2286     foreach my $commit ( @commits )
2287     {
2288         $self->{log}->debug("GITCVS::updater - Processing commit $commit->{hash} (" . (++$commitcount) . " of $committotal)");
2289         if (defined $lastpicked)
2290         {
2291             if (!in_array($lastpicked, @{$commit->{parents}}))
2292             {
2293                 # skip, we'll see this delta
2294                 # as part of a merge later
2295                 # warn "skipping off-track  $commit->{hash}\n";
2296                 next;
2297             } elsif (@{$commit->{parents}} > 1) {
2298                 # it is a merge commit, for each parent that is
2299                 # not $lastpicked, see if we can get a log
2300                 # from the merge-base to that parent to put it
2301                 # in the message as a merge summary.
2302                 my @parents = @{$commit->{parents}};
2303                 foreach my $parent (@parents) {
2304                     # git-merge-base can potentially (but rarely) throw
2305                     # several candidate merge bases. let's assume
2306                     # that the first one is the best one.
2307                     if ($parent eq $lastpicked) {
2308                         next;
2309                     }
2310                     open my $p, 'git-merge-base '. $lastpicked . ' '
2311                     . $parent . '|';
2312                     my @output = (<$p>);
2313                     close $p;
2314                     my $base = join('', @output);
2315                     chomp $base;
2316                     if ($base) {
2317                         my @merged;
2318                         # print "want to log between  $base $parent \n";
2319                         open(GITLOG, '-|', 'git-log', "$base..$parent")
2320                         or die "Cannot call git-log: $!";
2321                         my $mergedhash;
2322                         while (<GITLOG>) {
2323                             chomp;
2324                             if (!defined $mergedhash) {
2325                                 if (m/^commit\s+(.+)$/) {
2326                                     $mergedhash = $1;
2327                                 } else {
2328                                     next;
2329                                 }
2330                             } else {
2331                                 # grab the first line that looks non-rfc822
2332                                 # aka has content after leading space
2333                                 if (m/^\s+(\S.*)$/) {
2334                                     my $title = $1;
2335                                     $title = substr($title,0,100); # truncate
2336                                     unshift @merged, "$mergedhash $title";
2337                                     undef $mergedhash;
2338                                 }
2339                             }
2340                         }
2341                         close GITLOG;
2342                         if (@merged) {
2343                             $commit->{mergemsg} = $commit->{message};
2344                             $commit->{mergemsg} .= "\nSummary of merged commits:\n\n";
2345                             foreach my $summary (@merged) {
2346                                 $commit->{mergemsg} .= "\t$summary\n";
2347                             }
2348                             $commit->{mergemsg} .= "\n\n";
2349                             # print "Message for $commit->{hash} \n$commit->{mergemsg}";
2350                         }
2351                     }
2352                 }
2353             }
2354         }
2355
2356         # convert the date to CVS-happy format
2357         $commit->{date} = "$2 $1 $4 $3 $5" if ( $commit->{date} =~ /^\w+\s+(\w+)\s+(\d+)\s+(\d+:\d+:\d+)\s+(\d+)\s+([+-]\d+)$/ );
2358
2359         if ( defined ( $lastpicked ) )
2360         {
2361             my $filepipe = open(FILELIST, '-|', 'git-diff-tree', '-z', '-r', $lastpicked, $commit->{hash}) or die("Cannot call git-diff-tree : $!");
2362             local ($/) = "\0";
2363             while ( <FILELIST> )
2364             {
2365                 chomp;
2366                 unless ( /^:\d{6}\s+\d{3}(\d)\d{2}\s+[a-zA-Z0-9]{40}\s+([a-zA-Z0-9]{40})\s+(\w)$/o )
2367                 {
2368                     die("Couldn't process git-diff-tree line : $_");
2369                 }
2370                 my ($mode, $hash, $change) = ($1, $2, $3);
2371                 my $name = <FILELIST>;
2372                 chomp($name);
2373
2374                 # $log->debug("File mode=$mode, hash=$hash, change=$change, name=$name");
2375
2376                 my $git_perms = "";
2377                 $git_perms .= "r" if ( $mode & 4 );
2378                 $git_perms .= "w" if ( $mode & 2 );
2379                 $git_perms .= "x" if ( $mode & 1 );
2380                 $git_perms = "rw" if ( $git_perms eq "" );
2381
2382                 if ( $change eq "D" )
2383                 {
2384                     #$log->debug("DELETE   $name");
2385                     $head->{$name} = {
2386                         name => $name,
2387                         revision => $head->{$name}{revision} + 1,
2388                         filehash => "deleted",
2389                         commithash => $commit->{hash},
2390                         modified => $commit->{date},
2391                         author => $commit->{author},
2392                         mode => $git_perms,
2393                     };
2394                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2395                 }
2396                 elsif ( $change eq "M" )
2397                 {
2398                     #$log->debug("MODIFIED $name");
2399                     $head->{$name} = {
2400                         name => $name,
2401                         revision => $head->{$name}{revision} + 1,
2402                         filehash => $hash,
2403                         commithash => $commit->{hash},
2404                         modified => $commit->{date},
2405                         author => $commit->{author},
2406                         mode => $git_perms,
2407                     };
2408                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2409                 }
2410                 elsif ( $change eq "A" )
2411                 {
2412                     #$log->debug("ADDED    $name");
2413                     $head->{$name} = {
2414                         name => $name,
2415                         revision => 1,
2416                         filehash => $hash,
2417                         commithash => $commit->{hash},
2418                         modified => $commit->{date},
2419                         author => $commit->{author},
2420                         mode => $git_perms,
2421                     };
2422                     $self->insert_rev($name, $head->{$name}{revision}, $hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2423                 }
2424                 else
2425                 {
2426                     $log->warn("UNKNOWN FILE CHANGE mode=$mode, hash=$hash, change=$change, name=$name");
2427                     die;
2428                 }
2429             }
2430             close FILELIST;
2431         } else {
2432             # this is used to detect files removed from the repo
2433             my $seen_files = {};
2434
2435             my $filepipe = open(FILELIST, '-|', 'git-ls-tree', '-z', '-r', $commit->{hash}) or die("Cannot call git-ls-tree : $!");
2436             local $/ = "\0";
2437             while ( <FILELIST> )
2438             {
2439                 chomp;
2440                 unless ( /^(\d+)\s+(\w+)\s+([a-zA-Z0-9]+)\t(.*)$/o )
2441                 {
2442                     die("Couldn't process git-ls-tree line : $_");
2443                 }
2444
2445                 my ( $git_perms, $git_type, $git_hash, $git_filename ) = ( $1, $2, $3, $4 );
2446
2447                 $seen_files->{$git_filename} = 1;
2448
2449                 my ( $oldhash, $oldrevision, $oldmode ) = (
2450                     $head->{$git_filename}{filehash},
2451                     $head->{$git_filename}{revision},
2452                     $head->{$git_filename}{mode}
2453                 );
2454
2455                 if ( $git_perms =~ /^\d\d\d(\d)\d\d/o )
2456                 {
2457                     $git_perms = "";
2458                     $git_perms .= "r" if ( $1 & 4 );
2459                     $git_perms .= "w" if ( $1 & 2 );
2460                     $git_perms .= "x" if ( $1 & 1 );
2461                 } else {
2462                     $git_perms = "rw";
2463                 }
2464
2465                 # unless the file exists with the same hash, we need to update it ...
2466                 unless ( defined($oldhash) and $oldhash eq $git_hash and defined($oldmode) and $oldmode eq $git_perms )
2467                 {
2468                     my $newrevision = ( $oldrevision or 0 ) + 1;
2469
2470                     $head->{$git_filename} = {
2471                         name => $git_filename,
2472                         revision => $newrevision,
2473                         filehash => $git_hash,
2474                         commithash => $commit->{hash},
2475                         modified => $commit->{date},
2476                         author => $commit->{author},
2477                         mode => $git_perms,
2478                     };
2479
2480
2481                     $self->insert_rev($git_filename, $newrevision, $git_hash, $commit->{hash}, $commit->{date}, $commit->{author}, $git_perms);
2482                 }
2483             }
2484             close FILELIST;
2485
2486             # Detect deleted files
2487             foreach my $file ( keys %$head )
2488             {
2489                 unless ( exists $seen_files->{$file} or $head->{$file}{filehash} eq "deleted" )
2490                 {
2491                     $head->{$file}{revision}++;
2492                     $head->{$file}{filehash} = "deleted";
2493                     $head->{$file}{commithash} = $commit->{hash};
2494                     $head->{$file}{modified} = $commit->{date};
2495                     $head->{$file}{author} = $commit->{author};
2496
2497                     $self->insert_rev($file, $head->{$file}{revision}, $head->{$file}{filehash}, $commit->{hash}, $commit->{date}, $commit->{author}, $head->{$file}{mode});
2498                 }
2499             }
2500             # END : "Detect deleted files"
2501         }
2502
2503
2504         if (exists $commit->{mergemsg})
2505         {
2506             $self->insert_mergelog($commit->{hash}, $commit->{mergemsg});
2507         }
2508
2509         $lastpicked = $commit->{hash};
2510
2511         $self->_set_prop("last_commit", $commit->{hash});
2512     }
2513
2514     $self->delete_head();
2515     foreach my $file ( keys %$head )
2516     {
2517         $self->insert_head(
2518             $file,
2519             $head->{$file}{revision},
2520             $head->{$file}{filehash},
2521             $head->{$file}{commithash},
2522             $head->{$file}{modified},
2523             $head->{$file}{author},
2524             $head->{$file}{mode},
2525         );
2526     }
2527     # invalidate the gethead cache
2528     $self->{gethead_cache} = undef;
2529
2530
2531     # Ending exclusive lock here
2532     $self->{dbh}->commit() or die "Failed to commit changes to SQLite";
2533 }
2534
2535 sub insert_rev
2536 {
2537     my $self = shift;
2538     my $name = shift;
2539     my $revision = shift;
2540     my $filehash = shift;
2541     my $commithash = shift;
2542     my $modified = shift;
2543     my $author = shift;
2544     my $mode = shift;
2545
2546     my $insert_rev = $self->{dbh}->prepare_cached("INSERT INTO revision (name, revision, filehash, commithash, modified, author, mode) VALUES (?,?,?,?,?,?,?)",{},1);
2547     $insert_rev->execute($name, $revision, $filehash, $commithash, $modified, $author, $mode);
2548 }
2549
2550 sub insert_mergelog
2551 {
2552     my $self = shift;
2553     my $key = shift;
2554     my $value = shift;
2555
2556     my $insert_mergelog = $self->{dbh}->prepare_cached("INSERT INTO commitmsgs (key, value) VALUES (?,?)",{},1);
2557     $insert_mergelog->execute($key, $value);
2558 }
2559
2560 sub delete_head
2561 {
2562     my $self = shift;
2563
2564     my $delete_head = $self->{dbh}->prepare_cached("DELETE FROM head",{},1);
2565     $delete_head->execute();
2566 }
2567
2568 sub insert_head
2569 {
2570     my $self = shift;
2571     my $name = shift;
2572     my $revision = shift;
2573     my $filehash = shift;
2574     my $commithash = shift;
2575     my $modified = shift;
2576     my $author = shift;
2577     my $mode = shift;
2578
2579     my $insert_head = $self->{dbh}->prepare_cached("INSERT INTO head (name, revision, filehash, commithash, modified, author, mode) VALUES (?,?,?,?,?,?,?)",{},1);
2580     $insert_head->execute($name, $revision, $filehash, $commithash, $modified, $author, $mode);
2581 }
2582
2583 sub _headrev
2584 {
2585     my $self = shift;
2586     my $filename = shift;
2587
2588     my $db_query = $self->{dbh}->prepare_cached("SELECT filehash, revision, mode FROM head WHERE name=?",{},1);
2589     $db_query->execute($filename);
2590     my ( $hash, $revision, $mode ) = $db_query->fetchrow_array;
2591
2592     return ( $hash, $revision, $mode );
2593 }
2594
2595 sub _get_prop
2596 {
2597     my $self = shift;
2598     my $key = shift;
2599
2600     my $db_query = $self->{dbh}->prepare_cached("SELECT value FROM properties WHERE key=?",{},1);
2601     $db_query->execute($key);
2602     my ( $value ) = $db_query->fetchrow_array;
2603
2604     return $value;
2605 }
2606
2607 sub _set_prop
2608 {
2609     my $self = shift;
2610     my $key = shift;
2611     my $value = shift;
2612
2613     my $db_query = $self->{dbh}->prepare_cached("UPDATE properties SET value=? WHERE key=?",{},1);
2614     $db_query->execute($value, $key);
2615
2616     unless ( $db_query->rows )
2617     {
2618         $db_query = $self->{dbh}->prepare_cached("INSERT INTO properties (key, value) VALUES (?,?)",{},1);
2619         $db_query->execute($key, $value);
2620     }
2621
2622     return $value;
2623 }
2624
2625 =head2 gethead
2626
2627 =cut
2628
2629 sub gethead
2630 {
2631     my $self = shift;
2632
2633     return $self->{gethead_cache} if ( defined ( $self->{gethead_cache} ) );
2634
2635     my $db_query = $self->{dbh}->prepare_cached("SELECT name, filehash, mode, revision, modified, commithash, author FROM head ORDER BY name ASC",{},1);
2636     $db_query->execute();
2637
2638     my $tree = [];
2639     while ( my $file = $db_query->fetchrow_hashref )
2640     {
2641         push @$tree, $file;
2642     }
2643
2644     $self->{gethead_cache} = $tree;
2645
2646     return $tree;
2647 }
2648
2649 =head2 getlog
2650
2651 =cut
2652
2653 sub getlog
2654 {
2655     my $self = shift;
2656     my $filename = shift;
2657
2658     my $db_query = $self->{dbh}->prepare_cached("SELECT name, filehash, author, mode, revision, modified, commithash FROM revision WHERE name=? ORDER BY revision DESC",{},1);
2659     $db_query->execute($filename);
2660
2661     my $tree = [];
2662     while ( my $file = $db_query->fetchrow_hashref )
2663     {
2664         push @$tree, $file;
2665     }
2666
2667     return $tree;
2668 }
2669
2670 =head2 getmeta
2671
2672 This function takes a filename (with path) argument and returns a hashref of
2673 metadata for that file.
2674
2675 =cut
2676
2677 sub getmeta
2678 {
2679     my $self = shift;
2680     my $filename = shift;
2681     my $revision = shift;
2682
2683     my $db_query;
2684     if ( defined($revision) and $revision =~ /^\d+$/ )
2685     {
2686         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM revision WHERE name=? AND revision=?",{},1);
2687         $db_query->execute($filename, $revision);
2688     }
2689     elsif ( defined($revision) and $revision =~ /^[a-zA-Z0-9]{40}$/ )
2690     {
2691         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM revision WHERE name=? AND commithash=?",{},1);
2692         $db_query->execute($filename, $revision);
2693     } else {
2694         $db_query = $self->{dbh}->prepare_cached("SELECT * FROM head WHERE name=?",{},1);
2695         $db_query->execute($filename);
2696     }
2697
2698     return $db_query->fetchrow_hashref;
2699 }
2700
2701 =head2 commitmessage
2702
2703 this function takes a commithash and returns the commit message for that commit
2704
2705 =cut
2706 sub commitmessage
2707 {
2708     my $self = shift;
2709     my $commithash = shift;
2710
2711     die("Need commithash") unless ( defined($commithash) and $commithash =~ /^[a-zA-Z0-9]{40}$/ );
2712
2713     my $db_query;
2714     $db_query = $self->{dbh}->prepare_cached("SELECT value FROM commitmsgs WHERE key=?",{},1);
2715     $db_query->execute($commithash);
2716
2717     my ( $message ) = $db_query->fetchrow_array;
2718
2719     if ( defined ( $message ) )
2720     {
2721         $message .= " " if ( $message =~ /\n$/ );
2722         return $message;
2723     }
2724
2725     my @lines = safe_pipe_capture("git-cat-file", "commit", $commithash);
2726     shift @lines while ( $lines[0] =~ /\S/ );
2727     $message = join("",@lines);
2728     $message .= " " if ( $message =~ /\n$/ );
2729     return $message;
2730 }
2731
2732 =head2 gethistory
2733
2734 This function takes a filename (with path) argument and returns an arrayofarrays
2735 containing revision,filehash,commithash ordered by revision descending
2736
2737 =cut
2738 sub gethistory
2739 {
2740     my $self = shift;
2741     my $filename = shift;
2742
2743     my $db_query;
2744     $db_query = $self->{dbh}->prepare_cached("SELECT revision, filehash, commithash FROM revision WHERE name=? ORDER BY revision DESC",{},1);
2745     $db_query->execute($filename);
2746
2747     return $db_query->fetchall_arrayref;
2748 }
2749
2750 =head2 gethistorydense
2751
2752 This function takes a filename (with path) argument and returns an arrayofarrays
2753 containing revision,filehash,commithash ordered by revision descending.
2754
2755 This version of gethistory skips deleted entries -- so it is useful for annotate.
2756 The 'dense' part is a reference to a '--dense' option available for git-rev-list
2757 and other git tools that depend on it.
2758
2759 =cut
2760 sub gethistorydense
2761 {
2762     my $self = shift;
2763     my $filename = shift;
2764
2765     my $db_query;
2766     $db_query = $self->{dbh}->prepare_cached("SELECT revision, filehash, commithash FROM revision WHERE name=? AND filehash!='deleted' ORDER BY revision DESC",{},1);
2767     $db_query->execute($filename);
2768
2769     return $db_query->fetchall_arrayref;
2770 }
2771
2772 =head2 in_array()
2773
2774 from Array::PAT - mimics the in_array() function
2775 found in PHP. Yuck but works for small arrays.
2776
2777 =cut
2778 sub in_array
2779 {
2780     my ($check, @array) = @_;
2781     my $retval = 0;
2782     foreach my $test (@array){
2783         if($check eq $test){
2784             $retval =  1;
2785         }
2786     }
2787     return $retval;
2788 }
2789
2790 =head2 safe_pipe_capture
2791
2792 an alternative to `command` that allows input to be passed as an array
2793 to work around shell problems with weird characters in arguments
2794
2795 =cut
2796 sub safe_pipe_capture {
2797
2798     my @output;
2799
2800     if (my $pid = open my $child, '-|') {
2801         @output = (<$child>);
2802         close $child or die join(' ',@_).": $! $?";
2803     } else {
2804         exec(@_) or die "$! $?"; # exec() can fail the executable can't be found
2805     }
2806     return wantarray ? @output : join('',@output);
2807 }
2808
2809
2810 1;