git-cvsserver.perl

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