Archive.pm 31.8 KB
Newer Older
1
2
3
4
# -*- indent-tabs-mode: nil; -*-
# vim:ft=perl:et:sw=4
# $Id$

5
# Sympa - SYsteme de Multi-Postage Automatique
6
7
8
9
#
# Copyright (c) 1997, 1998, 1999 Institut Pasteur & Christophe Wolfhugel
# Copyright (c) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005,
# 2006, 2007, 2008, 2009, 2010, 2011 Comite Reseau des Universites
10
# Copyright (c) 2011, 2012, 2013, 2014, 2015, 2016, 2017 GIP RENATER
11
12
13
# Copyright 2018 The Sympa Community. See the AUTHORS.md file at the
# top-level directory of this distribution and at
# <https://github.com/sympa-community/sympa.git>.
14
15
16
17
18
19
20
21
22
23
24
25
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
26
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
root's avatar
root committed
27

28
package Sympa::Archive;
root's avatar
root committed
29

30
use strict;
31
use warnings;
sikeda's avatar
sikeda committed
32
use Cwd qw();
33
use Digest::MD5 qw();
sikeda's avatar
sikeda committed
34
use Encode qw();
35
36
use English qw(-no_match_vars);
use File::Path qw();
37
use IO::File;
38
use POSIX qw();
39

40
use Sympa;
41
use Conf;
42
43
use Sympa::Constants;
use Sympa::LockedFile;
44
use Sympa::Log;
45
use Sympa::Message;
46
use Sympa::Spool;
47
use Sympa::Tools::File;
48
use Sympa::Tools::Text;
root's avatar
root committed
49

50
51
my $log = Sympa::Log->instance;

52
53
sub new {
    my $class = shift;
54
55
56
57
58
59
60
61

    my $list;
    if (ref $_[0]) {    # Compat., not recommended.
        $list = shift;
    } else {
        my %options = @_;
        $list = $options{context};
    }
62
63
64
65
66

    die 'Bug in logic.  Ask developer' unless ref $list eq 'Sympa::List';

    my $self = bless {
        context           => $list,
67
        base_directory    => $list->get_archive_dir,
68
69
70
71
72
73
74
75
76
77
        arc_directory     => undef,
        directory         => undef,
        deleted_directory => undef,
        _metadatas        => undef,
    } => $class;

    $self->_create_spool();

    return $self;
}
root's avatar
root committed
78

79
80
81
82
83
84
85
86
87
88
89
90
91
92
sub _create_spool {
    my $self = shift;

    my $umask = umask oct $Conf::Conf{'umask'};
    foreach my $directory ($Conf::Conf{'arc_path'}, $self->{base_directory}) {
        unless (-d $directory) {
            $log->syslog('info', 'Creating spool %s', $directory);
            unless (
                mkdir($directory, 0755)
                and Sympa::Tools::File::set_file_rights(
                    file  => $directory,
                    user  => Sympa::Constants::USER(),
                    group => Sympa::Constants::GROUP()
                )
Luc Didry's avatar
Luc Didry committed
93
            ) {
94
95
96
97
98
99
                die sprintf 'Cannot create %s: %s', $directory, $ERRNO;
            }
        }
    }
    umask $umask;
}
root's avatar
root committed
100

101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
sub add_archive {
    my $self = shift;
    my $arc  = shift;

    return undef unless $arc;
    return undef unless $arc =~ /\A\d{4}-\d{2}\z/;

    my $umask = umask oct $Conf::Conf{'umask'};
    my $error;
    File::Path::make_path(
        $self->{base_directory} . '/' . $arc . '/arctxt',
        {   mode  => 0775,
            owner => Sympa::Constants::USER(),
            group => Sympa::Constants::GROUP(),
            error => \$error
        }
    );
    umask $umask;
root's avatar
root committed
119

120
121
122
123
124
    if (@$error) {
        return undef;
    }
    return 1;
}
125

126
127
128
sub purge_archive {
    my $self = shift;
    my $arc  = shift;
129

130
131
    return undef unless $arc;
    return undef unless $arc =~ /\A\d{4}-\d{2}\z/;
132

133
134
135
    my $error;
    File::Path::remove_tree($self->{base_directory} . '/' . $arc,
        {error => \$error});
136

137
138
    if (@$error) {
        return undef;
salaun's avatar
salaun committed
139
    }
140
141
    return 1;
}
142

143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
sub select_archive {
    my $self    = shift;
    my $arc     = shift;
    my %options = @_;

    return undef unless $arc;
    return undef unless $arc =~ /\A\d{4}-\d{2}\z/;

    my $arc_directory     = $self->{base_directory} . '/' . $arc;
    my $directory         = $arc_directory . '/arctxt';
    my $deleted_directory = $arc_directory . '/deleted';

    my $dh;
    return undef unless opendir $dh, $directory;
    closedir $dh;

    undef $self->{_metadatas};
sikeda's avatar
sikeda committed
160
    undef $self->{_html_metadatas};
161
162
163
164
165
166
    $self->{arc_directory}     = $arc_directory;
    $self->{directory}         = $directory;
    $self->{deleted_directory} = $deleted_directory;

    if ($options{info}) {
        return {
167
168
            size  => Sympa::Tools::File::get_dir_size($directory),
            mtime => Sympa::Tools::File::get_mtime($directory),
169
        };
sikeda's avatar
sikeda committed
170
171
172
173
174
175
176
177
    } elsif ($options{count}) {
        my $count;
        if (open my $fh, '<', $self->{arc_directory} . '/index') {
            $count = <$fh>;
            chomp $count;
            close $fh;
        }
        return {count => ($count || 0)};
178
179
180
    } else {
        return $arc;
    }
root's avatar
root committed
181
182
}

183
184
185
186
187
188
189
190
191
192
193
sub fetch {
    my $self    = shift;
    my %options = @_;

    undef $self->{_metadatas};    # Rewind cache.
    while (1) {
        my ($message, $handle) = $self->next;
        last unless $handle;      # No more messages.
        next unless $message;     # Malformed message.

        if ($options{message_id}) {
194
195
196
            my $message_id = Sympa::Tools::Text::canonic_message_id(
                $message->get_header('Message-Id'))
                || '';
197
198
199
200
201
202
203
204
205
            if ($message_id eq $options{message_id}) {
                undef $self->{_metadatas};    # Rewind cache.
                return ($message, $handle);
            }
        }
    }

    return;
}
sympa-authors's avatar
   
sympa-authors committed
206

sikeda's avatar
sikeda committed
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
sub html_fetch {
    $log->syslog('debug2', '(%s, %s => %s)', @_);
    my $self    = shift;
    my %options = @_;

    return undef unless $self->{arc_directory};
    return undef unless $options{file};

    my $handle =
        IO::File->new($self->{arc_directory} . '/' . $options{file}, '<');
    return undef unless $handle;

    my $metadata = {};    # May be empty.
    while (<$handle>) {
        last if /^\s*$/;    ## Metadata end with an emtpy line

        if (/^<!--(\S+): (.*) -->$/) {
            my ($key, $value) = ($1, $2);
225
            $value = Sympa::Tools::Text::decode_html($value);
sikeda's avatar
sikeda committed
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
            if ($key eq 'X-From-R13') {
                $metadata->{'X-From'} = $value;
                # MHonArc protection of email addresses.
                $metadata->{'X-From'} =~ tr/N-Z[@A-Mn-za-m/@A-Z[a-z/;
                # Remove the gecos.
                $metadata->{'X-From'} =~ s/^.*<(.*)>/$1/g;
            }
            $metadata->{$key} = $value;
        }
    }
    seek $handle, 0, 0;
    $metadata->{html_content} = do { local $RS; <$handle> };
    $metadata->{filename} = $options{file};

    return $metadata;
}

243
sub next {
244
245
    my $self    = shift;
    my %options = @_;
246

247
    return unless $self->{directory};
sympa-authors's avatar
   
sympa-authors committed
248

249
250
251
252
253
254
255
256
257
258
    unless ($self->{_metadatas}) {
        my $dh;
        unless (opendir $dh, $self->{directory}) {
            die sprintf 'Cannot open dir %s: %s', $self->{directory}, $ERRNO;
        }
        $self->{_metadatas} = [
            sort _cmp_numeric grep {
                        !/,lock/
                    and !m{(?:\A|/)(?:\.|T\.|BAD-)}
                    and -f ($self->{directory} . '/' . $_)
259
            } readdir $dh
260
261
        ];
        closedir $dh;
262
263
264
265

        # The "reverse" option specific to this class is set.
        $self->{_metadatas} = [reverse @{$self->{_metadatas}}]
            if $options{reverse};
266
    }
267
268
269
    unless (@{$self->{_metadatas}}) {
        undef $self->{_metadatas};
        return;
270
    }
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290

    while (my $marshalled = shift @{$self->{_metadatas}}) {
        my ($lock_fh, $metadata, $message);

        # Try locking message.  Those locked or removed by other process will
        # be skipped.
        $lock_fh =
            Sympa::LockedFile->new($self->{directory} . '/' . $marshalled,
            -1, '+<');
        next unless $lock_fh;

        $metadata =
            Sympa::Spool::unmarshal_metadata($self->{directory}, $marshalled,
            qr{\A(\d+)\z}, [qw(serial)]);

        if ($metadata) {
            my $msg_string = do { local $RS; <$lock_fh> };
            $message = Sympa::Message->new($msg_string, %$metadata);
        }

sikeda's avatar
sikeda committed
291
292
293
        # Metadata doesn't contain context; add it.
        $message->{context} = $self->{context} if $message;

294
295
        # Though message might not be deserialized, anyway return the result.
        return ($message, $lock_fh);
root's avatar
root committed
296
    }
297
    return;
root's avatar
root committed
298
299
}

sikeda's avatar
sikeda committed
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
sub html_next {
    my $self    = shift;
    my %options = @_;

    return undef unless $self->{arc_directory};

    unless ($self->{_html_metadatas}) {
        my $dh;
        unless (opendir $dh, $self->{arc_directory}) {
            die sprintf 'Cannot open dir %s: %s', $self->{arc_directory},
                $ERRNO;
        }
        $self->{_html_metadatas} = [
            sort _cmp_numeric grep {
                        !/,lock/
                    and !m{(?:\A|/)(?:\.|T\.|BAD-)}
                    and -f ($self->{arc_directory} . '/' . $_)
317
            } readdir $dh
sikeda's avatar
sikeda committed
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
        ];
        closedir $dh;

        # The "reverse" option specific to this class is set.
        $self->{_html_metadatas} = [reverse @{$self->{_html_metadatas}}]
            if $options{reverse};
    }
    unless (@{$self->{_html_metadatas}}) {
        undef $self->{_html_metadatas};
        return undef;
    }

    while (my $marshalled = shift @{$self->{_html_metadatas}}) {
        return $self->html_fetch(file => $marshalled);
    }
    return undef;
}

336
sub _cmp_numeric {
sikeda's avatar
sikeda committed
337
338
339
340
    my $a_num = $1 if defined $a and $a =~ /(\d+)/;
    my $b_num = $1 if defined $b and $b =~ /(\d+)/;
    if (defined $a_num and defined $b_num) {
        return $a_num <=> $b_num || $a cmp $b;
341
342
343
344
345
346
347
348
349
350
351
    } else {
        return $a cmp $b;
    }
}

sub remove {
    $log->syslog('debug2', '(%s, %s)', @_);
    my $self   = shift;
    my $handle = shift;

    return undef unless $self->{arc_directory};
352

353
    my $list = $self->{context};
354

355
356
357
358
359
360
361
362
363
364
365
366
367
368
    # Move text message to deleted/ directory.
    unless (-d $self->{deleted_directory}) {
        my $umask = umask oct $Conf::Conf{'umask'};
        unless (mkdir $self->{deleted_directory}, 0777) {
            die sprintf 'Unable to create %s: %s',
                $self->{deleted_directory}, $ERRNO;
        }
        umask $umask;
    }
    unless (
        $handle->rename($self->{deleted_directory} . '/' . $handle->basename))
    {
        $log->syslog('info', 'Unable to rename message %s in archive %s: %s',
            $handle->basename, $self, Sympa::LockedFile->last_error);
369
        return undef;
sympa-authors's avatar
   
sympa-authors committed
370
    }
371

372
373
    # Remove directory if empty arctxt.
    rmdir $self->{directory};
374

375
376
    return 1;
}
sympa-authors's avatar
   
sympa-authors committed
377

378
# Old name: remove() in archived.pl.
sikeda's avatar
sikeda committed
379
sub html_remove {
380
381
    my $self  = shift;
    my $msgid = shift;
sympa-authors's avatar
   
sympa-authors committed
382

383
384
    return undef unless $self->{arc_directory};
    return undef unless $msgid and $msgid !~ /NO-ID-FOUND\.mhonarc\.org/;
sympa-authors's avatar
   
sympa-authors committed
385

386
    my $list = $self->{context};
sympa-authors's avatar
   
sympa-authors committed
387

388
389
390
391
392
393
    # Remove message from HTML archive.
    system(
        Conf::get_robot_conf($list->{'domain'}, 'mhonarc'),
        '-outdir' => $self->{arc_directory},
        '-rmm'    => $msgid
    );
sympa-authors's avatar
   
sympa-authors committed
394

395
    # Remomve urlized message.
396
397
398
399
    my $url_dir =
          $list->{'dir'}
        . '/urlized/'
        . Sympa::Tools::Text::escape_chars($msgid);
400
401
    my $error;
    File::Path::remove_tree($url_dir, {error => \$error});
sympa-authors's avatar
   
sympa-authors committed
402

403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
    return 1;
}

# Does the real job : stores the message given as an argument into
# the indicated directory.
# Old name: (part of) mail2arc() in archived.pl.
sub store {
    my $self    = shift;
    my $message = shift;

    my $list = $self->{context};
    my $arc = POSIX::strftime('%Y-%m', localtime $message->{date});
    my $newfile;

    unless ($self->select_archive($arc)) {
        $self->add_archive($arc);
        unless ($self->select_archive($arc)) {
            $log->syslog('err', 'Cannot create directory %s in archive %s',
                $arc, $self);
            return undef;
        }
sympa-authors's avatar
   
sympa-authors committed
424
425
    }

426
427
428
429
430
431
432
433
434
435
436
437
438
    # Copy the file in the arctxt.
    if (-f $self->{arc_directory} . "/index") {
        open my $fh, '<', $self->{arc_directory} . '/index'
            or die sprintf 'Can\'t read index of %s in %s: %s', $arc, $self,
            $ERRNO;
        $newfile = <$fh>;
        chomp $newfile;
        $newfile++;
        close $fh;
    } else {
        # recreate index file if needed and update it
        $newfile = _create_idx($self->{arc_directory}) + 1;
    }
sympa-authors's avatar
   
sympa-authors committed
439

440
441
442
443
444
445
    # Save arctxt dump of original message.
    open my $fh, '>', $self->{directory} . '/' . $newfile
        or die sprintf 'Can\'t open file %s/%s: %s', $self->{directory},
        $newfile, $ERRNO;
    print $fh $message->as_string;
    close $fh;
sympa-authors's avatar
   
sympa-authors committed
446

447
    _save_idx($self->{arc_directory} . '/index', $newfile);
sympa-authors's avatar
   
sympa-authors committed
448

449
450
451
452
    $log->syslog('notice', 'Message %s is stored into archive %s as <%s>',
        $message, $self, $newfile);
    return $newfile;
}
453

454
# Old name: (part of) mail2arc in archived.pl.
sikeda's avatar
sikeda committed
455
sub html_store {
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
    my $self    = shift;
    my $message = shift->dup;

    my $list = $self->{context};
    my $arc  = POSIX::strftime('%Y-%m', localtime $message->{date});
    my $yyyy = POSIX::strftime('%Y', localtime $message->{date});
    my $mm   = POSIX::strftime('%m', localtime $message->{date});

    unless ($self->select_archive($arc)) {
        $self->add_archive($arc);
        unless ($self->select_archive($arc)) {
            $log->syslog('err', 'Cannot create directory %s in archive %s',
                $arc, $self);
            return undef;
        }
    }
472

473
474
475
    # Prepare clean message content (HTML parts are cleaned)
    unless ($message->clean_html) {
        $log->syslog('err', "Could not clean message, ignoring message");
476
477
        return undef;
    }
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494

    my $mhonarc_ressources =
        Sympa::search_fullpath($list, 'mhonarc-ressources.tt2');

    $log->syslog(
        'debug',
        'Calling %s for list %s',
        Conf::get_robot_conf($list->{'domain'}, 'mhonarc'), $list
    );

    my $tag = _get_tag($list);

    # Call mhonarc on cleaned message source to make clean htlm view of
    # message.
    my @cmd = (
        Conf::get_robot_conf($list->{'domain'}, 'mhonarc'),
        '-add',
sikeda's avatar
sikeda committed
495
496
497
498
        '-addressmodifycode' => '1',    # w/a: Clear old cache in .mhonarc.db.
        '-rcfile'     => $mhonarc_ressources,
        '-outdir'     => $self->{arc_directory},
        '-definevars' => sprintf(
499
            "listname='%s' hostname=%s yyyy=%s mois=%s yyyymm=%s-%s wdir=%s base=%s/arc tag=%s with_tslice=1 with_powered_by=1",
500
501
502
503
504
505
506
            $list->{'name'},
            $list->{'domain'},
            $yyyy,
            $mm,
            $yyyy,
            $mm,
            Conf::get_robot_conf($list->{'domain'}, 'arc_path'),
IKEDA Soji's avatar
IKEDA Soji committed
507
            (Conf::get_robot_conf($list->{'domain'}, 'wwsympa_url') || ''),
508
509
510
511
512
513
514
515
516
517
            $tag
        ),
        '-umask' => $Conf::Conf{'umask'}
    );

    $log->syslog('debug', 'System call: %s', join(' ', @cmd));

    my $pipeout;
    unless (open $pipeout, '|-', @cmd) {
        $log->syslog('err', 'Could not open pipe: %m');
518
        return undef;
sympa-authors's avatar
   
sympa-authors committed
519
    }
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
    print $pipeout $message->as_string;
    close $pipeout;
    my $status = $? >> 8;

    ## Remove lock if required
    if ($status == 75) {
        $log->syslog(
            'notice',
            'Removing lock directory %s',
            $self->{arc_directory} . '/.mhonarc.lck'
        );
        rmdir $self->{arc_directory} . '/.mhonarc.lck';

        my $pipeout;
        unless (open $pipeout, '|-', @cmd) {
            $log->syslog('err', 'Could not open pipe: %m');
            return undef;
537
        }
538
539
540
        print $pipeout $message->as_string;
        close $pipeout;
        $status = $? >> 8;
sympa-authors's avatar
   
sympa-authors committed
541
    }
542
543
544
545
546
547
548
549
550
    if ($status) {
        $log->syslog(
            'err',
            'Command %s failed with exit code %s',
            join(' ', @cmd), $status
        );
    }

    return 1;
sympa-authors's avatar
   
sympa-authors committed
551
552
}

553
554
# DEPRECATED.  No longer used.
#sub store_last;
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574

# DEPRECATED.  Use get_archives() and select_archive().
#sub list;

# Lists the files included in the archive, preformatted for printing
# Returns an array.
sub get_archives {
    $log->syslog('debug2', '(%s)', @_);
    my $self = shift;

    my $base_directory = $self->{base_directory};

    my $dh;
    unless ($base_directory and opendir $dh, $base_directory) {
        $log->syslog('err', 'Cannot open directory %s: %m', $base_directory);
        return;
    }
    my @arcs =
        grep {
        /\A\d\d\d\d\-\d\d\z/
575
            and -d $base_directory . '/'
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
            . $_
            . '/arctxt'
        }
        sort readdir $dh;
    closedir $dh;

    return @arcs;
}

# DEPRECATED.  Use select_archive() and next().
#sub scan_dir_archive;

# DEPRECATED.  Use select_archive() and fetch().
#sub search_msgid;

591
592
593
594
595
596
597
598
599
# Old name: Sympa::List::get_arc_size().
sub get_size {
    my $self = shift;
    my $dir  = shift;

    return 0 unless -d $self->{base_directory};
    return Sympa::Tools::File::get_dir_size($self->{base_directory});
}

600
# OBSOLETED.  No longer used.
root's avatar
root committed
601
sub exist {
602
    my ($name, $file) = @_;
root's avatar
root committed
603
    my $fn = "$name/$file";
604

root's avatar
root committed
605
606
607
608
    return $fn if (-r $fn && -f $fn);
    return undef;
}

sympa-authors's avatar
sympa-authors committed
609
# return path for latest message distributed in the list
610
611
# DEPRECATED.  No longer used.
#sub last_path;
612
613
614

## Load an archived message, returns the mhonarc metadata
## IN : file_path
sikeda's avatar
sikeda committed
615
616
# DEPRECATED.  Use html_fetch() or html_next().
#sub load_html_message;
617

618
# Old name: rebuild() in archived.pl.
sikeda's avatar
sikeda committed
619
sub html_rebuild {
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
    my $self = shift;
    my $arc  = shift;

    $arc =~ /^(\d{4})-(\d{2})$/;
    my $yyyy = $1;
    my $mm   = $2;

    return unless $self->select_archive($arc);

    my $list          = $self->{context};
    my $listname      = $list->{'name'};
    my $robot_id      = $list->{'domain'};
    my $arc_directory = $self->{arc_directory};

    my $tag = _get_tag($list);
    my $mhonarc_ressources =
        Sympa::search_fullpath($list, 'mhonarc-ressources.tt2');

    # Remove existing HTML files and .mhonarc.db.
    my $dh;
    opendir $dh, $arc_directory;
    unlink map { $arc_directory . '/' . $_ }
        grep {
                $_ ne 'arctxt'
            and $_ ne 'index'
            and $_ ne 'deleted'
            and !/\A\.+\z/
        } readdir $dh;
    closedir $dh;

    my $dir_to_rebuild = $self->{directory};
    my $arcs_dir       = $self->_clean_archive_directory($arc);
    if ($arcs_dir) {
        $dir_to_rebuild = $arcs_dir->{'dir_to_rebuild'};
    }

    # recreate index file if needed
    unless (-f $arc_directory . '/index') {
        _create_idx($arc_directory);
    }

    my @cmd = (
        Conf::get_robot_conf($robot_id, 'mhonarc'),
sikeda's avatar
sikeda committed
663
664
665
666
        '-addressmodifycode' => '1',    # w/a: Clear old cache in .mhonarc.db.
        '-rcfile'     => $mhonarc_ressources,
        '-outdir'     => $arc_directory,
        '-definevars' => sprintf(
667
            "listname='%s' hostname=%s yyyy=%s mois=%s yyyymm=%s-%s wdir=%s base=%s/arc tag=%s with_tslice=1 with_powered_by=1",
668
669
670
671
672
673
674
            $listname,
            $robot_id,
            $yyyy,
            $mm,
            $yyyy,
            $mm,
            Conf::get_robot_conf($robot_id, 'arc_path'),
IKEDA Soji's avatar
IKEDA Soji committed
675
            (Conf::get_robot_conf($robot_id, 'wwsympa_url') || ''),
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
            $tag
        ),
        '-umask' => $Conf::Conf{'umask'},
        $dir_to_rebuild
    );
    my $exitcode = system(@cmd) >> 8;

    # Delete temporary directory containing files with escaped HTML.
    if ($arcs_dir and -d $arcs_dir->{'cleaned_dir'}) {
        my $error;
        File::Path::remove_tree($arcs_dir->{'cleaned_dir'},
            {error => \$error});
    }

    ## Remove lock if required
    if ($exitcode == 75) {
        $log->syslog(
            'notice',
            'Removing lock directory %s',
            $arc_directory . '/.mhonarc.lck'
        );
        rmdir $arc_directory . '/.mhonarc.lck';

        $exitcode = system(@cmd) >> 8;
    }
    if ($exitcode) {
        $log->syslog(
            'err',
            'Command %s failed with exit code %s',
            join(' ', @cmd), $exitcode
        );
    }
}

# Sets the value of $ENV{'M2H_ADDRESSMODIFYCODE'} and
# $ENV{'M2H_MODIFYBODYADDRESSES'}.
#* $tag a character string (containing the result of _get_tag($list))
713
# NO LONGER USED.
sikeda's avatar
sikeda committed
714
#sub _set_hidden_mode;
715
716

# Empties $ENV{'M2H_ADDRESSMODIFYCODE'}.
717
# NO LONGER USED.
sikeda's avatar
sikeda committed
718
#sub _unset_hidden_mode;
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762

# Saves the archives index file
#* $index, a string corresponding to the file name to which save an index.
#* $lst, a character string
# Old name: save_idx() in archived.pl.
sub _save_idx {
    my ($index, $lst) = @_;

    return unless $lst;

    open INDEXF, '>', $index
        or die sprintf 'Couldn\'t overwrite index %s: %s', $index, $!;
    print INDEXF "$lst\n";
    close INDEXF;
}

# Create the 'index' file for one archive subdir
# Old name: create_idx() in archived.pl.
sub _create_idx {
    my $arc_dir = shift;    ## corresponds to the yyyy-mm directory

    my $arc_txt_dir = $arc_dir . '/arctxt';

    unless (opendir(DIR, $arc_txt_dir)) {
        $log->syslog('err', 'Failed to open directory "%s"', $arc_txt_dir);
        return undef;
    }

    my @files = (sort { $a <=> $b; } grep(/^\d+$/, (readdir DIR)));
    my $index = $files[$#files] || 0;
    _save_idx($arc_dir . '/index', $index);

    closedir DIR;

    return $index;
}

# Old name: clean_archive_directory().
sub _clean_archive_directory {
    $log->syslog('debug3', '(%s, %s)', @_);
    my $self = shift;
    my $arc  = shift;

    return undef unless $self->select_archive($arc);
763
764

    my $answer;
765
766
767
    $answer->{'dir_to_rebuild'} = $self->{directory};
    $answer->{'cleaned_dir'}    = sprintf '%s/%s/%s/arctxt',
        $Conf::Conf{'tmpdir'}, $self->{context}->get_id, $arc;
768
    unless (
769
770
771
        my $number_of_copies = Sympa::Tools::File::copy_dir(
            $answer->{'dir_to_rebuild'},
            $answer->{'cleaned_dir'}
772
        )
Luc Didry's avatar
Luc Didry committed
773
    ) {
774
        $log->syslog(
775
            'err',
776
            'Unable to create a temporary directory where to store files for HTML escaping (%s). Cancelling',
777
778
779
780
781
782
783
784
785
            $number_of_copies
        );
        return undef;
    }
    if (opendir ARCDIR, $answer->{'cleaned_dir'}) {
        my $files_left_uncleaned = 0;
        foreach my $file (readdir(ARCDIR)) {
            next if ($file =~ /^\./);
            $files_left_uncleaned++
786
787
                unless _clean_archived_message(
                $self->{context}->{'domain'},    #FIXME
788
789
790
                $answer->{'cleaned_dir'} . '/' . $file,
                $answer->{'cleaned_dir'} . '/' . $file
                );
791
        }
792
        closedir ARCDIR;
793
        if ($files_left_uncleaned) {
794
            $log->syslog('err',
795
                'HTML cleaning failed for %s files in the directory %s',
796
797
798
799
                $files_left_uncleaned, $answer->{'dir_to_rebuild'});
        }
        $answer->{'dir_to_rebuild'} = $answer->{'cleaned_dir'};
    } else {
800
        $log->syslog(
801
            'err',
802
803
            'Unable to open directory %s: %m',
            $answer->{'dir_to_rebuild'}
804
        );
805
        Sympa::Tools::File::del_dir($answer->{'cleaned_dir'});
806
807
808
809
        return undef;
    }
    return $answer;
}
810

811
812
# Old name: clean_archived_message().
sub _clean_archived_message {
813
    $log->syslog('debug2', '(%s, %s, %s)', @_);
814
815
    my $robot  = shift;
    my $input  = shift;
816
817
    my $output = shift;

818
    my $message = Sympa::Message->new_from_file($input, context => $robot);
819
    unless ($message) {
820
        $log->syslog('err', 'Unable to create a Message object with file %s',
821
822
823
            $input);
        return undef;
    }
824

825
    if ($message->clean_html) {
826
827
        if (open TMP, '>', $output) {
            print TMP $message->as_string;
828
829
830
            close TMP;
            return 1;
        } else {
831
            $log->syslog(
832
                'err',
833
                'Unable to create a tmp file to write clean HTML to file %s',
834
835
                $output
            );
836
837
838
            return undef;
        }
    } else {
839
        $log->syslog('err', 'HTML cleaning in file %s failed', $output);
840
        return undef;
841
842
    }
}
root's avatar
root committed
843

sikeda's avatar
sikeda committed
844
845
846
847
# Old names archive::convert_single_msg_2_html(),
# Sympa::Archive::convert_single_message().
sub html_format {
    my $message = shift;
848
    my %opts    = @_;
849

sikeda's avatar
sikeda committed
850
    my $that = $message->{context};
851
852
853
    my $list;
    my $robot;
    my $listname;
854
    my $domain;
855
    if (ref $that eq 'Sympa::List') {
856
857
858
        $list     = $that;
        $robot    = $that->{'domain'};
        $listname = $that->{'name'};
859
        $domain   = $that->{'domain'};
860
861
862
863
    } elsif (!ref($that) and $that and $that ne '*') {
        $list     = '';
        $robot    = $that;
        $listname = '';
864
        $domain   = Conf::get_robot_conf($that, 'domain');
865
    } else {
866
        die 'bug in logic.  Ask developer';
867
868
    }

sikeda's avatar
sikeda committed
869
    my $msg_as_string = $message->as_string;
root's avatar
root committed
870

871
    my $destination_dir = $opts{'destination_dir'};
872
    my $attachment_url  = $opts{'attachment_url'};
873
874
875
876
    if (ref $attachment_url eq 'ARRAY') {
        $attachment_url = join '/',
            map { Sympa::Tools::Text::encode_uri($_) } @$attachment_url;
    }
877

878
    my $mhonarc_ressources =
879
        Sympa::search_fullpath($that, 'mhonarc-ressources.tt2');
880
    unless ($mhonarc_ressources) {
881
        $log->syslog('notice', 'Cannot find any MhOnArc ressource file');
882
        return undef;
883
884
885
    }

    unless (-d $destination_dir) {
886
        unless (Sympa::Tools::File::mkdir_all($destination_dir, 0755)) {
887
            $log->syslog('err', 'Unable to create %s', $destination_dir);
888
889
            return undef;
        }
890
    }
891
892
893

    my $msg_file = $destination_dir . '/msg00000.txt';
    unless (open OUT, '>', $msg_file) {
sikeda's avatar
sikeda committed
894
        $log->syslog('notice', 'Can\'t open %s', $msg_file);
895
        return undef;
896
    }
897
898
899
    print OUT $msg_as_string;
    close OUT;

900
    # mhonarc require du change workdir so this proc must retore it
sikeda's avatar
sikeda committed
901
    my $pwd = Cwd::getcwd();
902

903
904
    ## generate HTML
    unless (chdir $destination_dir) {
905
        $log->syslog('err', 'Could not change working directory to %s',
906
907
            $destination_dir);
        return undef;
908
909
    }

910
    my $tag      = _get_tag($that);
911
    my $exitcode = system(
sikeda's avatar
sikeda committed
912
913
914
        Conf::get_robot_conf($robot, 'mhonarc'),
        '-single',
        '-rcfile'     => $mhonarc_ressources,
915
        '-definevars' => sprintf(
916
            "listname='%s' hostname=%s yyyy='' mois='' tag=%s with_tslice='' with_powered_by=''",
917
            $listname, $domain, $tag
sikeda's avatar
sikeda committed
918
        ),
919
        '-outdir'        => $destination_dir,
920
        '-attachmentdir' => $destination_dir,
921
922
923
924
        '-attachmenturl' =>
            sprintf('(%s%% path_cgi %%%s)/%s', $tag, $tag, $attachment_url),
        '-umask'  => $Conf::Conf{'umask'},
        '-stdout' => "$destination_dir/msg00000.html",
sikeda's avatar
sikeda committed
925
926
        '--',
        $msg_file
927
    ) >> 8;
928

929
930
    # restore current wd
    chdir $pwd;
931

932
    if ($exitcode) {
933
        $log->syslog(
934
935
936
937
            'err',
            'Command %s failed with exit code %d',
            Conf::get_robot_conf($robot, 'mhonarc'), $exitcode
        );
938
939
    }

940
941
942
    return 1;
}

943
944
945
# Old name: Sympa::Archive::get_tag(), get_tag() in archived.pl.
sub _get_tag {
    my $that = shift;
946

947
948
949
950
951
952
953
954
    my $name;
    if (ref $that eq 'Sympa::List') {
        $name = $that->{'name'};
    } elsif (!ref($that) and $that and $that ne '*') {
        $name = $that;
    } elsif (!ref($that)) {
        $name = '*';
    }
955

956
957
958
959
    my $cookie = $Conf::Conf{'cookie'};
    $cookie = '' unless defined $cookie;
    return substr(Digest::MD5::md5_hex(join '/', $cookie, $name), -10);
}
960

961
962
sub get_id {
    my $self = shift;
963

964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
    my $context = $self->{context};
    unless (ref $context eq 'Sympa::List') {
        return '';
    } elsif ($self->{arc_directory}) {
        return sprintf '%s/%s', $context->get_id,
            [split '/', $self->{arc_directory}]->[-1];
    } else {
        return $context->get_id;
    }
}

1;
__END__

=encoding utf-8

=head1 NAME

Sympa::Archive - Archives of Sympa

=head1 SYNOPSIS

  use Sympa::Archive;
987
  $archive = Sympa::Archive->new(context => $list);
988
989
990
991

  @arcs = $archive->get_archives;

  $archive->store($message);
sikeda's avatar
sikeda committed
992
  $archive->html_store($message);
993
994
995
996
997
998

  $archive->select_archive('2015-04');
  ($message, $handle) = $archive->next;

  $archive->select_archive('2015-04');
  ($message, $handle) = $archive->fetch(message_id => $message_id);
sikeda's avatar
sikeda committed
999
  $archive->html_remove($message_id);
1000
  $archive->remove($handle);