summaryrefslogtreecommitdiffstats
path: root/summarize_time.cgi
blob: 26cc047257b11e53d9ee73837b86e92a746ff7e3 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
#!/usr/bin/perl -wT
# -*- Mode: perl; indent-tabs-mode: nil -*-
#
# The contents of this file are subject to the Mozilla Public
# License Version 1.1 (the "License"); you may not use this file
# except in compliance with the License. You may obtain a copy of
# the License at http://www.mozilla.org/MPL/
#
# Software distributed under the License is distributed on an "AS
# IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or
# implied. See the License for the specific language governing
# rights and limitations under the License.
#
# The Original Code is the Bugzilla Bug Tracking System.
#
# Contributor(s): Christian Reis <kiko@async.com.br>
#                 Shane H. W. Travis <travis@sedsystems.ca>
#
use strict;

use lib qw(.);

use Date::Parse;         # strptime
use Date::Format;        # strftime

use Bugzilla;
use Bugzilla::Constants; # LOGIN_*
use Bugzilla::Bug;       # EmitDependList
use Bugzilla::Util;      # trim
use Bugzilla::Error;
use Bugzilla::User;      # Bugzilla->user->in_group

my $template = Bugzilla->template;
my $vars = {};

#
# Date handling
#

sub date_adjust_down {
   
    my ($year, $month, $day) = @_;

    if ($day == 0) {
        $month -= 1;
        $day = 31;
        # Proper day adjustment is done later.

        if ($month == 0) {
            $year -= 1;
            $month = 12;
        }
    }

    if (($month == 2) && ($day > 28)) {
        if ($year % 4 == 0 && $year % 100 != 0) {
            $day = 29;
        } else {
            $day = 28;
        }
    }

    if (($month == 4 || $month == 6 || $month == 9 || $month == 11) &&
        ($day == 31) ) 
    {
        $day = 30;
    }
    return ($year, $month, $day);
}

sub date_adjust_up {
    my ($year, $month, $day) = @_;

    if ($day > 31) {
        $month += 1;
        $day    = 1;

        if ($month == 13) {
            $month = 1;
            $year += 1;
        }
    }

    if ($month == 2 && $day > 28) {
        if ($year % 4 != 0 || $year % 100 == 0 || $day > 29) {
            $month = 3;
            $day = 1;
        }
    }

    if (($month == 4 || $month == 6 || $month == 9 || $month == 11) &&
        ($day == 31) )
    {
        $month += 1; 
        $day    = 1;
    }

    return ($year, $month, $day);
}

sub check_dates {
    my ($start_date, $end_date) = @_;
    if ($start_date) {
        if (!str2time($start_date)) {
            ThrowUserError("illegal_date", {'date' => $start_date});
        }
        # This code may strike you as funny. It's actually a workaround
        # for an "issue" in str2time. If you enter the date 2004-06-31,
        # even though it's a bogus date (there *are* only 30 days in
        # June), it will parse and return 2004-07-01. To make this
        # less painful to the end-user, I do the "normalization" here,
        # but it might be "surprising" and warrant a warning in the end.
        $start_date = time2str("%Y-%m-%d", str2time($start_date));
    } 
    if ($end_date) {
        if (!str2time($end_date)) {
            ThrowUserError("illegal_date", {'date' => $end_date});
        }
        # see related comment above.
        $end_date = time2str("%Y-%m-%d", str2time($end_date));
    }
    return ($start_date, $end_date);
}

sub split_by_month {
    # Takes start and end dates and splits them into a list of
    # monthly-spaced 2-lists of dates.
    my ($start_date, $end_date) = @_;

    # We assume at this point that the dates are provided and sane
    my (undef, undef, undef, $sd, $sm, $sy, undef) = strptime($start_date);
    my (undef, undef, undef, $ed, $em, $ey, undef) = strptime($end_date);

    # Find out how many months fit between the two dates so we know
    # how many times we loop.
    my $yd = $ey - $sy;
    my $md = 12 * $yd + $em - $sm;
    # If the end day is smaller than the start day, last interval is not a whole month.
    if ($sd > $ed) {
        $md -= 1;
    }

    my (@months, $sub_start, $sub_end);
    # This +1 and +1900 are a result of strptime's bizarre semantics
    my $year = $sy + 1900;
    my $month = $sm + 1;

    # Keep the original date, when the date will be changed in the adjust_date.
    my $sd_tmp = $sd;
    my $month_tmp = $month;
    my $year_tmp = $year;

    # This section handles only the whole months.
    for (my $i=0; $i < $md; $i++) {
        # Start of interval is adjusted up: 31.2. -> 1.3.
        ($year_tmp, $month_tmp, $sd_tmp) = date_adjust_up($year, $month, $sd);
        $sub_start = sprintf("%04d-%02d-%02d", $year_tmp, $month_tmp, $sd_tmp); 
        $month += 1;
        if ($month == 13) {
            $month = 1;
            $year += 1;
        }
        # End of interval is adjusted down: 31.2 -> 28.2.
        ($year_tmp, $month_tmp, $sd_tmp) = date_adjust_down($year, $month, $sd - 1);
        $sub_end = sprintf("%04d-%02d-%02d", $year_tmp, $month_tmp, $sd_tmp);
        push @months, [$sub_start, $sub_end];
    }
    
    # This section handles the last (unfinished) month. 
    $sub_end = sprintf("%04d-%02d-%02d", $ey + 1900, $em + 1, $ed);
    ($year_tmp, $month_tmp, $sd_tmp) = date_adjust_up($year, $month, $sd);
    $sub_start = sprintf("%04d-%02d-%02d", $year_tmp, $month_tmp, $sd_tmp);
    push @months, [$sub_start, $sub_end];

    return @months;
}

sub include_tt_details {
    my ($res, $bugids, $start_date, $end_date) = @_;


    my $dbh = Bugzilla->dbh;
    my ($date_bits, $date_values) = sqlize_dates($start_date, $end_date);
    my $buglist = join ", ", @{$bugids};

    my $q = qq{SELECT bugs.bug_id, profiles.login_name, bugs.deadline,
                      bugs.estimated_time, bugs.remaining_time
               FROM   longdescs
               INNER JOIN bugs
                  ON longdescs.bug_id = bugs.bug_id
               INNER JOIN profiles
                  ON longdescs.who = profiles.userid
               WHERE  longdescs.bug_id in ($buglist) $date_bits};

    my %res = %{$res};
    my $sth = $dbh->prepare($q);
    $sth->execute(@{$date_values});
    while (my $row = $sth->fetch) {
        $res{$row->[0]}{"deadline"} = $row->[2];
        $res{$row->[0]}{"estimated_time"} = $row->[3];
        $res{$row->[0]}{"remaining_time"} = $row->[4];
    }
    return \%res;
}

sub sqlize_dates {
    my ($start_date, $end_date) = @_;
    my $date_bits = "";
    my @date_values;
    if ($start_date) {
        # we've checked, trick_taint is fine
        trick_taint($start_date);
        $date_bits = " AND longdescs.bug_when > ?";
        push @date_values, $start_date;
    } 
    if ($end_date) {
        # we need to add one day to end_date to catch stuff done today
        # do not forget to adjust date if it was the last day of month
        my (undef, undef, undef, $ed, $em, $ey, undef) = strptime($end_date);
        ($ey, $em, $ed) = date_adjust_up($ey+1900, $em+1, $ed+1);
        $end_date = sprintf("%04d-%02d-%02d", $ey, $em, $ed);

        $date_bits .= " AND longdescs.bug_when < ?"; 
        push @date_values, $end_date;
    }
    return ($date_bits, \@date_values);
}

#
# Dependencies
#

sub get_blocker_ids_unique {
    my $bug_id = shift;
    my @ret = ($bug_id);
    get_blocker_ids_deep($bug_id, \@ret);
    my %unique;
    foreach my $blocker (@ret) {
        $unique{$blocker} = $blocker
    }
    return keys %unique;
}

sub get_blocker_ids_deep {
    my ($bug_id, $ret) = @_;
    my $deps = Bugzilla::Bug::EmitDependList("blocked", "dependson", $bug_id);
    push @{$ret}, @$deps;
    foreach $bug_id (@$deps) {
        get_blocker_ids_deep($bug_id, $ret);
    }
}

#
# Queries and data structure assembly
#

sub query_work_by_buglist {
    my ($bugids, $start_date, $end_date) = @_;
    my $dbh = Bugzilla->dbh;

    my ($date_bits, $date_values) = sqlize_dates($start_date, $end_date);

    # $bugids is guaranteed to be non-empty because at least one bug is
    # always provided to this page.
    my $buglist = join ", ", @{$bugids};

    # Returns the total time worked on each bug *per developer*, with
    # bug descriptions and developer address
    my $q = qq{SELECT sum(longdescs.work_time) as total_time,
                      profiles.login_name, 
                      longdescs.bug_id,
                      bugs.short_desc,
                      bugs.bug_status
               FROM   longdescs
               INNER JOIN profiles
                   ON longdescs.who = profiles.userid
               INNER JOIN bugs
                   ON bugs.bug_id = longdescs.bug_id
               WHERE  longdescs.bug_id IN ($buglist)
                      $date_bits } .
            $dbh->sql_group_by('longdescs.bug_id, profiles.login_name',
                'bugs.short_desc, bugs.bug_status, longdescs.bug_when') . qq{
               ORDER BY longdescs.bug_when};
    my $sth = $dbh->prepare($q);
    $sth->execute(@{$date_values});
    return $sth;
}

sub get_work_by_owners {
    my $sth = query_work_by_buglist(@_);
    my %res;
    while (my $row = $sth->fetch) {
        # XXX: Why do we need to check if the total time is positive
        # instead of using SQL to do that?  Simply because MySQL 3.x's
        # GROUP BY doesn't work correctly with aggregates. This is
        # really annoying, but I've spent a long time trying to wrestle
        # with it and it just doesn't seem to work. Should work OK in
        # 4.x, though.
        if ($row->[0] > 0) {
            my $login_name = $row->[1];
            push @{$res{$login_name}}, { total_time => $row->[0],
                                         bug_id     => $row->[2],
                                         short_desc => $row->[3],
                                         bug_status => $row->[4] };
        }
    }
    return \%res;
}

sub get_work_by_bugs {
    my $sth = query_work_by_buglist(@_);
    my %res;
    while (my $row = $sth->fetch) {
        # Perl doesn't let me use arrays as keys :-(
        # merge in ID, status and summary
        my $bug = join ";", ($row->[2], $row->[4], $row->[3]);
        # XXX: see comment in get_work_by_owners
        if ($row->[0] > 0) {
            push @{$res{$bug}}, { total_time => $row->[0],
                                  login_name => $row->[1], };
        }
    }
    return \%res;
}

sub get_inactive_bugs {
    my ($bugids, $start_date, $end_date) = @_;
    my $dbh = Bugzilla->dbh;
    my ($date_bits, $date_values) = sqlize_dates($start_date, $end_date);
    my $buglist = join ", ", @{$bugids};

    my %res;
    # This sucks. I need to make sure that even bugs that *don't* show
    # up in the longdescs query (because no comments were filed during
    # the specified period) but *are* dependent on the parent bug show
    # up in the results if they have no work done; that's why I prefill
    # them in %res here and then remove them below.
    my $q = qq{SELECT DISTINCT bugs.bug_id, bugs.short_desc ,
                               bugs.bug_status
               FROM   longdescs
               INNER JOIN bugs
                    ON longdescs.bug_id = bugs.bug_id
               WHERE  longdescs.bug_id in ($buglist)};
    my $sth = $dbh->prepare($q);
    $sth->execute();
    while (my $row = $sth->fetch) {
        $res{$row->[0]} = [$row->[1], $row->[2]];
    }

    # Returns the total time worked on each bug, with description. This
    # query differs a bit from one in the query_work_by_buglist and I
    # avoided complicating that one just to make it more general.
    $q = qq{SELECT sum(longdescs.work_time) as total_time,
                   longdescs.bug_id,
                   bugs.short_desc,
                   bugs.bug_status
            FROM   longdescs
            INNER JOIN bugs
                ON bugs.bug_id = longdescs.bug_id 
            WHERE  longdescs.bug_id IN ($buglist)
                   $date_bits } .
         $dbh->sql_group_by('longdescs.bug_id',
                            'bugs.short_desc, bugs.bug_status,
                             longdescs.bug_when') . qq{
            ORDER BY longdescs.bug_when};
    $sth = $dbh->prepare($q);
    $sth->execute(@{$date_values});
    while (my $row = $sth->fetch) {
        # XXX: see comment in get_work_by_owners
        if ($row->[0] == 0) {
            $res{$row->[1]} = [$row->[2], $row->[3]];
        } else {
            delete $res{$row->[1]};
        }
    }
    return \%res;
}

#
# Misc
#

sub sort_bug_keys {
    # XXX a hack is the mother of all evils. The fact that we store keys
    # joined by semi-colons in the workdata-by-bug structure forces us to
    # write this evil comparison function to ensure we can process the
    # data timely -- just pushing it through a numerical sort makes TT
    # hang while generating output :-(
    my $list = shift;
    my @a;
    my @b;
    return sort { @a = split(";", $a); 
                  @b = split(";", $b); 
                  $a[0] <=> $b[0] } @{$list};
}

#
# Template code starts here
#

Bugzilla->login(LOGIN_REQUIRED);

my $cgi = Bugzilla->cgi;

Bugzilla->switch_to_shadow_db();

Bugzilla->user->in_group(Bugzilla->params->{"timetrackinggroup"})
    || ThrowUserError("auth_failure", {group  => "time-tracking",
                                       action => "access",
                                       object => "timetracking_summaries"});

my @ids = split(",", $cgi->param('id'));
map { ValidateBugID($_) } @ids;
@ids = map { detaint_natural($_) && $_ } @ids;
@ids = grep { Bugzilla->user->can_see_bug($_) } @ids;

my $group_by = $cgi->param('group_by') || "number";
my $monthly = $cgi->param('monthly');
my $detailed = $cgi->param('detailed');
my $do_report = $cgi->param('do_report');
my $inactive = $cgi->param('inactive');
my $do_depends = $cgi->param('do_depends');
my $ctype = scalar($cgi->param("ctype"));

my ($start_date, $end_date);
if ($do_report && @ids) {
    my @bugs = @ids;

    # Dependency mode requires a single bug and grabs dependents.
    if ($do_depends) {
        if (scalar(@bugs) != 1) {
            ThrowCodeError("bad_arg", { argument=>"id",
                                        function=>"summarize_time"});
        }
        @bugs = get_blocker_ids_unique($bugs[0]);
        @bugs = grep { Bugzilla->user->can_see_bug($_) } @bugs;
    }

    $start_date = trim $cgi->param('start_date');
    $end_date = trim $cgi->param('end_date');

    # Swap dates in case the user put an end_date before the start_date
    if ($start_date && $end_date && 
        str2time($start_date) > str2time($end_date)) {
        $vars->{'warn_swap_dates'} = 1;
        ($start_date, $end_date) = ($end_date, $start_date);
    }
    ($start_date, $end_date) = check_dates($start_date, $end_date);

    if ($detailed) {
        my %detail_data;
        my $res = include_tt_details(\%detail_data, \@bugs, $start_date, $end_date);

        $vars->{'detail_data'} = $res;
    }
  
    # Store dates ia session cookie the dates so re-visiting the page
    # for other bugs keeps them around.
    $cgi->send_cookie(-name => 'time-summary-dates',
                      -value => join ";", ($start_date, $end_date));

    my (@parts, $part_data, @part_list);

    # Break dates apart into months if necessary; if not, we use the
    # same @parts list to allow us to use a common codepath.
    if ($monthly) {
        # unfortunately it's not too easy to guess a start date, since
        # it depends on what bugs we're looking at. We risk bothering
        # the user here. XXX: perhaps run a query to see what the
        # earliest activity in longdescs for all bugs and use that as a
        # start date.
        $start_date || ThrowUserError("illegal_date", {'date' => $start_date});
        # we can, however, provide a default end date. Note that this
        # differs in semantics from the open-ended queries we use when
        # start/end_date aren't provided -- and clock skews will make
        # this evident!
        @parts = split_by_month($start_date, 
                                $end_date || time2str("%Y-%m-%d", time()));
    } else {
        @parts = ([$start_date, $end_date]);
    }

    my %empty_hash;
    # For each of the separate divisions, grab the relevant summaries 
    foreach my $part (@parts) {
        my ($sub_start, $sub_end) = @{$part};
        if (@bugs) {
            if ($group_by eq "owner") {
                $part_data = get_work_by_owners(\@bugs, $sub_start, $sub_end);
            } else {
                $part_data = get_work_by_bugs(\@bugs, $sub_start, $sub_end);
            }
        } else {
            # $part_data must be a reference to a hash
            $part_data = \%empty_hash; 
        }
        push @part_list, $part_data;
    }

    if ($inactive && @bugs) {
        $vars->{'null'} = get_inactive_bugs(\@bugs, $start_date, $end_date);
    } else {
        $vars->{'null'} = \%empty_hash;
    }

    $vars->{'part_list'} = \@part_list;
    $vars->{'parts'} = \@parts;

} elsif ($cgi->cookie("time-summary-dates")) {
    ($start_date, $end_date) = split ";", $cgi->cookie('time-summary-dates');
}

$vars->{'ids'} = \@ids;
$vars->{'start_date'} = $start_date;
$vars->{'end_date'} = $end_date;
$vars->{'group_by'} = $group_by;
$vars->{'monthly'} = $monthly;
$vars->{'detailed'} = $detailed;
$vars->{'inactive'} = $inactive;
$vars->{'do_report'} = $do_report;
$vars->{'do_depends'} = $do_depends;
$vars->{'check_time'} = \&check_time;
$vars->{'sort_bug_keys'} = \&sort_bug_keys;

my $format = $template->get_format("bug/summarize-time", undef, $ctype);

# Get the proper content-type
print $cgi->header(-type=> $format->{'ctype'});
$template->process("$format->{'template'}", $vars)
  || ThrowTemplateError($template->error());