line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
#!/usr/bin/perl -w |
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
# |
4
|
|
|
|
|
|
|
# dbmerge.pm |
5
|
|
|
|
|
|
|
# Copyright (C) 1991-2015 by John Heidemann |
6
|
|
|
|
|
|
|
# $Id: b578ea71b5b2a54bed1fd517633261c2832cd066 $ |
7
|
|
|
|
|
|
|
# |
8
|
|
|
|
|
|
|
# This program is distributed under terms of the GNU general |
9
|
|
|
|
|
|
|
# public license, version 2. See the file COPYING |
10
|
|
|
|
|
|
|
# in $dblibdir for details. |
11
|
|
|
|
|
|
|
# |
12
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
package Fsdb::Filter::dbmerge; |
14
|
|
|
|
|
|
|
|
15
|
|
|
|
|
|
|
=head1 NAME |
16
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
dbmerge - merge all inputs in sorted order based on the the specified columns |
18
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
=head1 SYNOPSIS |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
dbmerge --input A.fsdb --input B.fsdb [-T TemporaryDirectory] [-nNrR] column [column...] |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
or |
24
|
|
|
|
|
|
|
cat A.fsdb | dbmerge --input - --input B.fsdb [-T TemporaryDirectory] [-nNrR] column [column...] |
25
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
or |
28
|
|
|
|
|
|
|
dbmerge [-T TemporaryDirectory] [-nNrR] column [column...] --inputs A.fsdb [B.fsdb ...] |
29
|
|
|
|
|
|
|
|
30
|
|
|
|
|
|
|
|
31
|
|
|
|
|
|
|
=head1 DESCRIPTION |
32
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
Merge all provided, pre-sorted input files, producing one sorted result. |
34
|
|
|
|
|
|
|
Inputs can both be specified with C<--input>, or one can come |
35
|
|
|
|
|
|
|
from standard input and the other from C<--input>. |
36
|
|
|
|
|
|
|
With C<--xargs>, each line of standard input is a filename for input. |
37
|
|
|
|
|
|
|
|
38
|
|
|
|
|
|
|
Inputs must have identical schemas (columns, column order, |
39
|
|
|
|
|
|
|
and field separators). |
40
|
|
|
|
|
|
|
|
41
|
|
|
|
|
|
|
Unlike F, F supports an arbitrary number of |
42
|
|
|
|
|
|
|
input files. |
43
|
|
|
|
|
|
|
|
44
|
|
|
|
|
|
|
Because this program is intended to merge multiple sources, |
45
|
|
|
|
|
|
|
it does I default to reading from standard input. |
46
|
|
|
|
|
|
|
If you wish to list F<-> as an explicit input source. |
47
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
Also, because we deal with multiple input files, |
49
|
|
|
|
|
|
|
this module doesn't output anything until it's run. |
50
|
|
|
|
|
|
|
|
51
|
|
|
|
|
|
|
L consumes a fixed amount of memory regardless of input size. |
52
|
|
|
|
|
|
|
It therefore buffers output on disk as necessary. |
53
|
|
|
|
|
|
|
(Merging is implemented a series of two-way merges, |
54
|
|
|
|
|
|
|
so disk space is O(number of records).) |
55
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
L will merge data in parallel, if possible. |
57
|
|
|
|
|
|
|
The <--parallelism> option can control the degree of parallelism, |
58
|
|
|
|
|
|
|
if desired. |
59
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
|
61
|
|
|
|
|
|
|
=head1 OPTIONS |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
General option: |
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
=over 4 |
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
=item B<--xargs> |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
Expect that input filenames are given, one-per-line, on standard input. |
70
|
|
|
|
|
|
|
(In this case, merging can start incrementally. |
71
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
=item B<--removeinputs> |
73
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
Delete the source files after they have been consumed. |
75
|
|
|
|
|
|
|
(Defaults off, leaving the inputs in place.) |
76
|
|
|
|
|
|
|
|
77
|
|
|
|
|
|
|
=item B<-T TmpDir> |
78
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
where to put tmp files. |
80
|
|
|
|
|
|
|
Also uses environment variable TMPDIR, if -T is |
81
|
|
|
|
|
|
|
not specified. |
82
|
|
|
|
|
|
|
Default is /tmp. |
83
|
|
|
|
|
|
|
|
84
|
|
|
|
|
|
|
=item B<--parallelism N> or B<-j N> |
85
|
|
|
|
|
|
|
|
86
|
|
|
|
|
|
|
Allow up to N merges to happen in parallel. |
87
|
|
|
|
|
|
|
Default is the number of CPUs in the machine. |
88
|
|
|
|
|
|
|
|
89
|
|
|
|
|
|
|
=item B<--endgame> (or B<--noendgame>) |
90
|
|
|
|
|
|
|
|
91
|
|
|
|
|
|
|
Enable endgame mode, extra parallelism when finishing up. |
92
|
|
|
|
|
|
|
(On by default.) |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
=back |
95
|
|
|
|
|
|
|
|
96
|
|
|
|
|
|
|
Sort specification options (can be interspersed with column names): |
97
|
|
|
|
|
|
|
|
98
|
|
|
|
|
|
|
=over 4 |
99
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
=item B<-r> or B<--descending> |
101
|
|
|
|
|
|
|
|
102
|
|
|
|
|
|
|
sort in reverse order (high to low) |
103
|
|
|
|
|
|
|
|
104
|
|
|
|
|
|
|
=item B<-R> or B<--ascending> |
105
|
|
|
|
|
|
|
|
106
|
|
|
|
|
|
|
sort in normal order (low to high) |
107
|
|
|
|
|
|
|
|
108
|
|
|
|
|
|
|
=item B<-n> or B<--numeric> |
109
|
|
|
|
|
|
|
|
110
|
|
|
|
|
|
|
sort numerically |
111
|
|
|
|
|
|
|
|
112
|
|
|
|
|
|
|
=item B<-N> or B<--lexical> |
113
|
|
|
|
|
|
|
|
114
|
|
|
|
|
|
|
sort lexicographically |
115
|
|
|
|
|
|
|
|
116
|
|
|
|
|
|
|
=back |
117
|
|
|
|
|
|
|
|
118
|
|
|
|
|
|
|
=for comment |
119
|
|
|
|
|
|
|
begin_standard_fsdb_options |
120
|
|
|
|
|
|
|
|
121
|
|
|
|
|
|
|
This module also supports the standard fsdb options: |
122
|
|
|
|
|
|
|
|
123
|
|
|
|
|
|
|
=over 4 |
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
=item B<-d> |
126
|
|
|
|
|
|
|
|
127
|
|
|
|
|
|
|
Enable debugging output. |
128
|
|
|
|
|
|
|
|
129
|
|
|
|
|
|
|
=item B<-i> or B<--input> InputSource |
130
|
|
|
|
|
|
|
|
131
|
|
|
|
|
|
|
Read from InputSource, typically a file name, or C<-> for standard input, |
132
|
|
|
|
|
|
|
or (if in Perl) a IO::Handle, Fsdb::IO or Fsdb::BoundedQueue objects. |
133
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
=item B<-o> or B<--output> OutputDestination |
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
Write to OutputDestination, typically a file name, or C<-> for standard output, |
137
|
|
|
|
|
|
|
or (if in Perl) a IO::Handle, Fsdb::IO or Fsdb::BoundedQueue objects. |
138
|
|
|
|
|
|
|
|
139
|
|
|
|
|
|
|
=item B<--autorun> or B<--noautorun> |
140
|
|
|
|
|
|
|
|
141
|
|
|
|
|
|
|
By default, programs process automatically, |
142
|
|
|
|
|
|
|
but Fsdb::Filter objects in Perl do not run until you invoke |
143
|
|
|
|
|
|
|
the run() method. |
144
|
|
|
|
|
|
|
The C<--(no)autorun> option controls that behavior within Perl. |
145
|
|
|
|
|
|
|
|
146
|
|
|
|
|
|
|
=item B<--help> |
147
|
|
|
|
|
|
|
|
148
|
|
|
|
|
|
|
Show help. |
149
|
|
|
|
|
|
|
|
150
|
|
|
|
|
|
|
=item B<--man> |
151
|
|
|
|
|
|
|
|
152
|
|
|
|
|
|
|
Show full manual. |
153
|
|
|
|
|
|
|
|
154
|
|
|
|
|
|
|
=back |
155
|
|
|
|
|
|
|
|
156
|
|
|
|
|
|
|
=for comment |
157
|
|
|
|
|
|
|
end_standard_fsdb_options |
158
|
|
|
|
|
|
|
|
159
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
=head1 SAMPLE USAGE |
161
|
|
|
|
|
|
|
|
162
|
|
|
|
|
|
|
=head2 Input: |
163
|
|
|
|
|
|
|
|
164
|
|
|
|
|
|
|
File F: |
165
|
|
|
|
|
|
|
|
166
|
|
|
|
|
|
|
#fsdb cid cname |
167
|
|
|
|
|
|
|
11 numanal |
168
|
|
|
|
|
|
|
10 pascal |
169
|
|
|
|
|
|
|
|
170
|
|
|
|
|
|
|
File F: |
171
|
|
|
|
|
|
|
|
172
|
|
|
|
|
|
|
#fsdb cid cname |
173
|
|
|
|
|
|
|
12 os |
174
|
|
|
|
|
|
|
13 statistics |
175
|
|
|
|
|
|
|
|
176
|
|
|
|
|
|
|
These two files are both sorted by C, |
177
|
|
|
|
|
|
|
and they have identical schemas. |
178
|
|
|
|
|
|
|
|
179
|
|
|
|
|
|
|
=head2 Command: |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
dbmerge --input a.fsdb --input b.fsdb cname |
182
|
|
|
|
|
|
|
|
183
|
|
|
|
|
|
|
or |
184
|
|
|
|
|
|
|
|
185
|
|
|
|
|
|
|
cat a.fsdb | dbmerge --input b.fsdb cname |
186
|
|
|
|
|
|
|
|
187
|
|
|
|
|
|
|
=head2 Output: |
188
|
|
|
|
|
|
|
|
189
|
|
|
|
|
|
|
#fsdb cid cname |
190
|
|
|
|
|
|
|
11 numanal |
191
|
|
|
|
|
|
|
12 os |
192
|
|
|
|
|
|
|
10 pascal |
193
|
|
|
|
|
|
|
13 statistics |
194
|
|
|
|
|
|
|
# | dbmerge --input a.fsdb --input b.fsdb cname |
195
|
|
|
|
|
|
|
|
196
|
|
|
|
|
|
|
=head1 SEE ALSO |
197
|
|
|
|
|
|
|
|
198
|
|
|
|
|
|
|
L, |
199
|
|
|
|
|
|
|
L, |
200
|
|
|
|
|
|
|
L |
201
|
|
|
|
|
|
|
|
202
|
|
|
|
|
|
|
=head1 CLASS FUNCTIONS |
203
|
|
|
|
|
|
|
|
204
|
|
|
|
|
|
|
=cut |
205
|
|
|
|
|
|
|
|
206
|
|
|
|
|
|
|
|
207
|
|
|
|
|
|
|
@ISA = qw(Fsdb::Filter); |
208
|
|
|
|
|
|
|
($VERSION) = 2.0; |
209
|
|
|
|
|
|
|
|
210
|
1
|
|
|
1
|
|
28
|
use 5.010; |
|
1
|
|
|
|
|
4
|
|
211
|
1
|
|
|
1
|
|
7
|
use strict; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
34
|
|
212
|
1
|
|
|
1
|
|
7
|
use Carp qw(croak); |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
52
|
|
213
|
1
|
|
|
1
|
|
7
|
use Pod::Usage; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
86
|
|
214
|
|
|
|
|
|
|
|
215
|
1
|
|
|
1
|
|
359
|
use IO::Pipe; |
|
1
|
|
|
|
|
1319
|
|
|
1
|
|
|
|
|
36
|
|
216
|
1
|
|
|
1
|
|
331
|
use IO::Select; |
|
1
|
|
|
|
|
1717
|
|
|
1
|
|
|
|
|
55
|
|
217
|
|
|
|
|
|
|
|
218
|
1
|
|
|
1
|
|
15
|
use Fsdb::Filter; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
28
|
|
219
|
1
|
|
|
1
|
|
7
|
use Fsdb::Filter::dbmerge2; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
28
|
|
220
|
1
|
|
|
1
|
|
7
|
use Fsdb::IO::Reader; |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
24
|
|
221
|
1
|
|
|
1
|
|
6
|
use Fsdb::IO::Writer; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
23
|
|
222
|
1
|
|
|
1
|
|
11
|
use Fsdb::Support::NamedTmpfile; |
|
1
|
|
|
|
|
4
|
|
|
1
|
|
|
|
|
24
|
|
223
|
1
|
|
|
1
|
|
7
|
use Fsdb::Support::OS; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
40
|
|
224
|
1
|
|
|
1
|
|
6
|
use Fsdb::Support::Freds; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
4282
|
|
225
|
|
|
|
|
|
|
|
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
=head2 new |
228
|
|
|
|
|
|
|
|
229
|
|
|
|
|
|
|
$filter = new Fsdb::Filter::dbmerge(@arguments); |
230
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
Create a new object, taking command-line arguments. |
232
|
|
|
|
|
|
|
|
233
|
|
|
|
|
|
|
=cut |
234
|
|
|
|
|
|
|
|
235
|
|
|
|
|
|
|
sub new($@) { |
236
|
0
|
|
|
0
|
1
|
|
my $class = shift @_; |
237
|
0
|
|
|
|
|
|
my $self = $class->SUPER::new(@_); |
238
|
0
|
|
|
|
|
|
bless $self, $class; |
239
|
0
|
|
|
|
|
|
$self->set_defaults; |
240
|
0
|
|
|
|
|
|
$self->parse_options(@_); |
241
|
0
|
|
|
|
|
|
$self->SUPER::post_new(); |
242
|
0
|
|
|
|
|
|
return $self; |
243
|
|
|
|
|
|
|
} |
244
|
|
|
|
|
|
|
|
245
|
|
|
|
|
|
|
|
246
|
|
|
|
|
|
|
=head2 set_defaults |
247
|
|
|
|
|
|
|
|
248
|
|
|
|
|
|
|
$filter->set_defaults(); |
249
|
|
|
|
|
|
|
|
250
|
|
|
|
|
|
|
Internal: set up defaults. |
251
|
|
|
|
|
|
|
|
252
|
|
|
|
|
|
|
=cut |
253
|
|
|
|
|
|
|
|
254
|
|
|
|
|
|
|
sub set_defaults($) { |
255
|
0
|
|
|
0
|
1
|
|
my $self = shift @_; |
256
|
0
|
|
|
|
|
|
$self->SUPER::set_defaults(); |
257
|
0
|
|
|
|
|
|
$self->{_remove_inputs} = undef; |
258
|
0
|
|
|
|
|
|
$self->{_info}{input_count} = 2; |
259
|
0
|
|
|
|
|
|
$self->{_sort_argv} = []; |
260
|
0
|
|
|
|
|
|
$self->{_max_parallelism} = undef; |
261
|
0
|
|
|
|
|
|
$self->{_parallelism_available} = undef; |
262
|
0
|
|
|
|
|
|
$self->{_test} = ''; |
263
|
0
|
|
|
|
|
|
$self->{_xargs} = undef; |
264
|
0
|
|
|
|
|
|
$self->{_endgame} = 1; |
265
|
0
|
|
|
|
|
|
$self->set_default_tmpdir; |
266
|
|
|
|
|
|
|
} |
267
|
|
|
|
|
|
|
|
268
|
|
|
|
|
|
|
=head2 parse_options |
269
|
|
|
|
|
|
|
|
270
|
|
|
|
|
|
|
$filter->parse_options(@ARGV); |
271
|
|
|
|
|
|
|
|
272
|
|
|
|
|
|
|
Internal: parse command-line arguments. |
273
|
|
|
|
|
|
|
|
274
|
|
|
|
|
|
|
=cut |
275
|
|
|
|
|
|
|
|
276
|
|
|
|
|
|
|
sub parse_options($@) { |
277
|
0
|
|
|
0
|
1
|
|
my $self = shift @_; |
278
|
|
|
|
|
|
|
|
279
|
0
|
|
|
|
|
|
my(@argv) = @_; |
280
|
0
|
|
|
|
|
|
my $past_sort_options = undef; |
281
|
|
|
|
|
|
|
$self->get_options( |
282
|
|
|
|
|
|
|
\@argv, |
283
|
0
|
|
|
0
|
|
|
'help|?' => sub { pod2usage(1); }, |
284
|
0
|
|
|
0
|
|
|
'man' => sub { pod2usage(-verbose => 2); }, |
285
|
|
|
|
|
|
|
'autorun!' => \$self->{_autorun}, |
286
|
|
|
|
|
|
|
'close!' => \$self->{_close}, |
287
|
|
|
|
|
|
|
'd|debug+' => \$self->{_debug}, |
288
|
|
|
|
|
|
|
'endgame!' => \$self->{_endgame}, |
289
|
0
|
|
|
0
|
|
|
'i|input=s@' => sub { $self->parse_io_option('inputs', @_); }, |
290
|
|
|
|
|
|
|
'inputs!' => \$past_sort_options, |
291
|
|
|
|
|
|
|
'j|parallelism=i' => \$self->{_max_parallelism}, |
292
|
|
|
|
|
|
|
'log!' => \$self->{_logprog}, |
293
|
0
|
|
|
0
|
|
|
'o|output=s' => sub { $self->parse_io_option('output', @_); }, |
294
|
|
|
|
|
|
|
'removeinputs!' => \$self->{_remove_inputs}, |
295
|
|
|
|
|
|
|
'test=s' => \$self->{_test}, |
296
|
|
|
|
|
|
|
'T|tmpdir|tempdir=s' => \$self->{_tmpdir}, |
297
|
|
|
|
|
|
|
'xargs!' => \$self->{_xargs}, |
298
|
|
|
|
|
|
|
# sort key options: |
299
|
0
|
|
|
0
|
|
|
'n|numeric' => sub { $self->parse_sort_option(@_); }, |
300
|
0
|
|
|
0
|
|
|
'N|lexical' => sub { $self->parse_sort_option(@_); }, |
301
|
0
|
|
|
0
|
|
|
'r|descending' => sub { $self->parse_sort_option(@_); }, |
302
|
0
|
|
|
0
|
|
|
'R|ascending' => sub { $self->parse_sort_option(@_); }, |
303
|
0
|
0
|
|
0
|
|
|
'<>' => sub { if ($past_sort_options) { |
304
|
0
|
|
|
|
|
|
$self->parse_io_option('inputs', @_); |
305
|
|
|
|
|
|
|
} else { |
306
|
0
|
|
|
|
|
|
$self->parse_sort_option('<>', @_); |
307
|
|
|
|
|
|
|
}; |
308
|
|
|
|
|
|
|
}, |
309
|
0
|
0
|
|
|
|
|
) or pod2usage(2); |
310
|
|
|
|
|
|
|
} |
311
|
|
|
|
|
|
|
|
312
|
|
|
|
|
|
|
=head2 _pretty_fn |
313
|
|
|
|
|
|
|
|
314
|
|
|
|
|
|
|
_pretty_fn($fn) |
315
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
Internal: pretty-print a filename or Fsdb::BoundedQueue. |
317
|
|
|
|
|
|
|
|
318
|
|
|
|
|
|
|
=cut |
319
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
sub _pretty_fn { |
321
|
0
|
|
|
0
|
|
|
my($fn) = @_; |
322
|
0
|
0
|
|
|
|
|
return ref($fn) if (ref($fn)); |
323
|
0
|
|
|
|
|
|
return $fn; |
324
|
|
|
|
|
|
|
} |
325
|
|
|
|
|
|
|
|
326
|
|
|
|
|
|
|
|
327
|
|
|
|
|
|
|
=head2 segment_next_output |
328
|
|
|
|
|
|
|
|
329
|
|
|
|
|
|
|
$out = $self->segment_next_output($output_type) |
330
|
|
|
|
|
|
|
|
331
|
|
|
|
|
|
|
Internal: return a Fsdb::IO::Writer as $OUT |
332
|
|
|
|
|
|
|
that either points to our output or a temporary file, |
333
|
|
|
|
|
|
|
depending on how things are going. |
334
|
|
|
|
|
|
|
|
335
|
|
|
|
|
|
|
The $OUTPUT_TYPE can be 'final' or 'ipc' or 'file'. |
336
|
|
|
|
|
|
|
|
337
|
|
|
|
|
|
|
=cut |
338
|
|
|
|
|
|
|
|
339
|
|
|
|
|
|
|
sub segment_next_output($$) { |
340
|
0
|
|
|
0
|
1
|
|
my ($self, $output_type) = @_; |
341
|
0
|
|
|
|
|
|
my $out; |
342
|
0
|
0
|
|
|
|
|
if ($output_type eq 'final') { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
343
|
|
|
|
|
|
|
# $self->finish_io_option('output', -clone => $self->{_two_ins}[0]); |
344
|
|
|
|
|
|
|
# $out = $self->{_out}; |
345
|
0
|
|
|
|
|
|
$out = $self->{_output}; # will pass this to the dbmerge2 module |
346
|
0
|
0
|
|
|
|
|
print "# final output\n" if ($self->{_debug}); |
347
|
|
|
|
|
|
|
} elsif ($output_type eq 'file') { |
348
|
|
|
|
|
|
|
# dump to a file for merging |
349
|
0
|
|
|
|
|
|
my $tmpfile = Fsdb::Support::NamedTmpfile::alloc($self->{_tmpdir}); |
350
|
0
|
|
|
|
|
|
$self->{_files_cleanup}{$tmpfile} = 'NamedTmpfile'; |
351
|
0
|
|
|
|
|
|
$out = $tmpfile; # just return the name |
352
|
|
|
|
|
|
|
} elsif ($output_type eq 'ipc') { |
353
|
|
|
|
|
|
|
# endgame-mode: send stuff down in-memory queues |
354
|
|
|
|
|
|
|
# $out = shared_clone(new Fsdb::BoundedQueue); |
355
|
0
|
|
|
|
|
|
my $read_fh = new IO::Handle; |
356
|
0
|
|
|
|
|
|
my $write_fh = new IO::Handle; |
357
|
0
|
|
|
|
|
|
pipe $read_fh, $write_fh; |
358
|
|
|
|
|
|
|
# Without these lines we suffer mojobake on unicode |
359
|
|
|
|
|
|
|
# as shown in test base TEST/dbsort_merge_unicode.cmd. |
360
|
0
|
|
|
|
|
|
binmode $read_fh, ":utf8"; |
361
|
0
|
|
|
|
|
|
binmode $write_fh, ":utf8"; |
362
|
0
|
|
|
|
|
|
$out = [ $read_fh, $write_fh ]; |
363
|
|
|
|
|
|
|
} else { |
364
|
0
|
|
|
|
|
|
die "internal error: dbmege.pm:segment_next_output bad output_type: $output_type\n"; |
365
|
|
|
|
|
|
|
}; |
366
|
0
|
|
|
|
|
|
return $out; |
367
|
|
|
|
|
|
|
} |
368
|
|
|
|
|
|
|
|
369
|
|
|
|
|
|
|
|
370
|
|
|
|
|
|
|
=head2 segment_cleanup |
371
|
|
|
|
|
|
|
|
372
|
|
|
|
|
|
|
$out = $self->segment_cleanup($file); |
373
|
|
|
|
|
|
|
|
374
|
|
|
|
|
|
|
Internal: Clean up a file, if necessary. |
375
|
|
|
|
|
|
|
(Sigh, used to be function pointers, but |
376
|
|
|
|
|
|
|
not clear how they would interact with threads.) |
377
|
|
|
|
|
|
|
|
378
|
|
|
|
|
|
|
=cut |
379
|
|
|
|
|
|
|
|
380
|
|
|
|
|
|
|
sub segment_cleanup($$) { |
381
|
0
|
|
|
0
|
1
|
|
my($self, $file) = @_; |
382
|
0
|
0
|
|
|
|
|
if (ref($file)) { |
383
|
0
|
0
|
|
|
|
|
if (ref($file) =~ /^IO::/) { |
|
|
0
|
|
|
|
|
|
384
|
0
|
0
|
|
|
|
|
print "# closing IO::*\n" if ($self->{_debug}); |
385
|
0
|
|
|
|
|
|
$file->close; |
386
|
|
|
|
|
|
|
} elsif (ref($file) eq 'Fsdb::BoundedQueue') { |
387
|
|
|
|
|
|
|
# nothing to do |
388
|
|
|
|
|
|
|
} else { |
389
|
0
|
|
|
|
|
|
die "internal error: unknown type in dbmerge::segment_cleanup\n"; |
390
|
|
|
|
|
|
|
}; |
391
|
0
|
|
|
|
|
|
return; |
392
|
|
|
|
|
|
|
}; |
393
|
0
|
|
|
|
|
|
my($cleanup_type) = $self->{_files_cleanup}{$file}; |
394
|
0
|
0
|
|
|
|
|
die "bad (empty) file in dbmerge::segment_cleanup\n" |
395
|
|
|
|
|
|
|
if (!defined($file)); |
396
|
0
|
0
|
|
|
|
|
if (!defined($cleanup_type)) { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
397
|
0
|
0
|
|
|
|
|
print "# dbmerge: segment_cleanup: no cleanup for " . _pretty_fn($file) . "\n" if ($self->{_debug}); |
398
|
|
|
|
|
|
|
# nothing |
399
|
|
|
|
|
|
|
} elsif ($cleanup_type eq 'unlink') { |
400
|
0
|
0
|
|
|
|
|
print "# dbmerge: segment_cleanup: cleaning up $file\n" if ($self->{_debug}); |
401
|
0
|
|
|
|
|
|
unlink($file); |
402
|
|
|
|
|
|
|
} elsif ($cleanup_type eq 'NamedTmpfile') { |
403
|
0
|
0
|
|
|
|
|
print "# dbmerge: segment_cleanup: NamedTmpfile::cleanup_one $file\n" if ($self->{_debug}); |
404
|
0
|
|
|
|
|
|
Fsdb::Support::NamedTmpfile::cleanup_one($file); |
405
|
|
|
|
|
|
|
} else { |
406
|
0
|
|
|
|
|
|
die $self->{_prog} . ": internal error, unknown segment_cleanup type $cleanup_type\n"; |
407
|
|
|
|
|
|
|
}; |
408
|
|
|
|
|
|
|
} |
409
|
|
|
|
|
|
|
|
410
|
|
|
|
|
|
|
=head2 _unique_id |
411
|
|
|
|
|
|
|
|
412
|
|
|
|
|
|
|
$id = $self->_unique_id() |
413
|
|
|
|
|
|
|
|
414
|
|
|
|
|
|
|
Generate a sequence number for debugging. |
415
|
|
|
|
|
|
|
|
416
|
|
|
|
|
|
|
=cut |
417
|
|
|
|
|
|
|
sub _unique_id() { |
418
|
0
|
|
|
0
|
|
|
my($self) = @_; |
419
|
0
|
|
0
|
|
|
|
$self->{_unique_id} //= 0; |
420
|
0
|
|
|
|
|
|
return $self->{_unique_id}++; |
421
|
|
|
|
|
|
|
}; |
422
|
|
|
|
|
|
|
|
423
|
|
|
|
|
|
|
|
424
|
|
|
|
|
|
|
=head2 segments_merge2_run |
425
|
|
|
|
|
|
|
|
426
|
|
|
|
|
|
|
$out = $self->segments_merge2_run($out_fn, $is_final_output, |
427
|
|
|
|
|
|
|
$in0, $in1); |
428
|
|
|
|
|
|
|
|
429
|
|
|
|
|
|
|
|
430
|
|
|
|
|
|
|
Internal: do the actual merge2 work (maybe our parent put us in a |
431
|
|
|
|
|
|
|
thread, maybe not). |
432
|
|
|
|
|
|
|
|
433
|
|
|
|
|
|
|
=cut |
434
|
|
|
|
|
|
|
sub segments_merge2_run($$$$$$) { |
435
|
0
|
|
|
0
|
1
|
|
my($self, $out_fn, $is_final_output, $in0, $in1, $id) = @_; |
436
|
|
|
|
|
|
|
|
437
|
0
|
|
|
|
|
|
my @merge_options = qw(--autorun --nolog); |
438
|
|
|
|
|
|
|
push(@merge_options, '--noclose', '--saveoutput' => \$self->{_out}) |
439
|
0
|
0
|
|
|
|
|
if ($is_final_output); |
440
|
|
|
|
|
|
|
|
441
|
0
|
|
|
|
|
|
my $debug_msg = ''; |
442
|
0
|
0
|
|
|
|
|
if ($self->{_debug}) { |
443
|
0
|
0
|
|
|
|
|
$debug_msg = "(id $id) " . _pretty_fn($in0) . " with " . _pretty_fn($in1) . " to " . _pretty_fn($out_fn) . " " . ($is_final_output ? " (final)" : "") . " " . join(" ", @merge_options); |
444
|
|
|
|
|
|
|
}; |
445
|
|
|
|
|
|
|
print "# segments_merge2_run: merge start $debug_msg\n" |
446
|
0
|
0
|
|
|
|
|
if ($self->{_debug}); |
447
|
|
|
|
|
|
|
new Fsdb::Filter::dbmerge2(@merge_options, |
448
|
|
|
|
|
|
|
'--input' => $in0, |
449
|
|
|
|
|
|
|
'--input' => $in1, |
450
|
|
|
|
|
|
|
'--output' => $out_fn, |
451
|
0
|
|
|
|
|
|
@{$self->{_sort_argv}}); |
|
0
|
|
|
|
|
|
|
452
|
|
|
|
|
|
|
print "# segments_merge2_run: merge finish $debug_msg\n" |
453
|
0
|
0
|
|
|
|
|
if ($self->{_debug}); |
454
|
|
|
|
|
|
|
|
455
|
0
|
|
|
|
|
|
$self->segment_cleanup($in0); |
456
|
0
|
|
|
|
|
|
$self->segment_cleanup($in1); |
457
|
0
|
0
|
|
|
|
|
if (!$is_final_output) { |
458
|
|
|
|
|
|
|
print "# segments_merge2_run: merge closing out " . ref($out_fn) . " $debug_msg\n" |
459
|
0
|
0
|
|
|
|
|
if ($self->{_debug}); |
460
|
|
|
|
|
|
|
}; |
461
|
|
|
|
|
|
|
} |
462
|
|
|
|
|
|
|
|
463
|
|
|
|
|
|
|
|
464
|
|
|
|
|
|
|
=head2 enqueue_work |
465
|
|
|
|
|
|
|
|
466
|
|
|
|
|
|
|
$self->enqueue_work($depth, $work); |
467
|
|
|
|
|
|
|
|
468
|
|
|
|
|
|
|
Internal: put $WORK on the queue at $DEPTH, updating the max count. |
469
|
|
|
|
|
|
|
|
470
|
|
|
|
|
|
|
=cut |
471
|
|
|
|
|
|
|
|
472
|
|
|
|
|
|
|
sub enqueue_work($$$) { |
473
|
0
|
|
|
0
|
1
|
|
my($self, $depth, $work) = @_; |
474
|
0
|
|
0
|
|
|
|
$self->{_work_max_files}[$depth] //= 0; |
475
|
0
|
|
|
|
|
|
$self->{_work_max_files}[$depth]++; |
476
|
0
|
|
|
|
|
|
push(@{$self->{_work}[$depth]}, $work); |
|
0
|
|
|
|
|
|
|
477
|
|
|
|
|
|
|
}; |
478
|
|
|
|
|
|
|
|
479
|
|
|
|
|
|
|
=head2 segments_merge_one_depth |
480
|
|
|
|
|
|
|
|
481
|
|
|
|
|
|
|
$self->segments_merge_one_depth($depth); |
482
|
|
|
|
|
|
|
|
483
|
|
|
|
|
|
|
Merge queued files, if any. |
484
|
|
|
|
|
|
|
|
485
|
|
|
|
|
|
|
Also release any queued threads. |
486
|
|
|
|
|
|
|
|
487
|
|
|
|
|
|
|
=cut |
488
|
|
|
|
|
|
|
|
489
|
|
|
|
|
|
|
sub segments_merge_one_depth($$) { |
490
|
0
|
|
|
0
|
1
|
|
my($self, $depth) = @_; |
491
|
|
|
|
|
|
|
|
492
|
0
|
|
|
|
|
|
my $work_depth_ref = $self->{_work}[$depth]; |
493
|
0
|
0
|
|
|
|
|
return if ($#$work_depth_ref == -1); # no work at this dpeth for now |
494
|
|
|
|
|
|
|
|
495
|
0
|
|
|
|
|
|
my $closed = $self->{_work_closed}[$depth]; |
496
|
|
|
|
|
|
|
|
497
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: scanning $depth\n" if ($self->{_debug}); |
498
|
|
|
|
|
|
|
# |
499
|
|
|
|
|
|
|
# Merge the files in a binary tree. |
500
|
|
|
|
|
|
|
# |
501
|
|
|
|
|
|
|
# In the past, we did this in a very clever |
502
|
|
|
|
|
|
|
# a file-system-cache-friendly order, based on ideas from |
503
|
|
|
|
|
|
|
# "Information and Control in Gray-box Systems" by |
504
|
|
|
|
|
|
|
# the Arpaci-Dusseaus at SOSP 2001. |
505
|
|
|
|
|
|
|
# |
506
|
|
|
|
|
|
|
# Unfortunately, this optimization makes the sort unstable |
507
|
|
|
|
|
|
|
# and complicated, |
508
|
|
|
|
|
|
|
# so it was dropped when paralleism was added. |
509
|
|
|
|
|
|
|
# |
510
|
0
|
0
|
|
|
|
|
while ($#{$work_depth_ref} >= ($closed ? 0 : 3)) { |
|
0
|
|
|
|
|
|
|
511
|
0
|
0
|
|
|
|
|
if ($#{$work_depth_ref} == 0) { |
|
0
|
|
|
|
|
|
|
512
|
0
|
0
|
|
|
|
|
last if (!$closed); |
513
|
|
|
|
|
|
|
# one left, just punt it next |
514
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: runt at depth $depth pushed to next depth.\n" if ($self->{_debug}); |
515
|
0
|
|
|
|
|
|
$self->enqueue_work($depth + 1, shift @{$work_depth_ref}); |
|
0
|
|
|
|
|
|
|
516
|
0
|
0
|
|
|
|
|
die "internal error\n" if ($#{$work_depth_ref} != -1); |
|
0
|
|
|
|
|
|
|
517
|
0
|
|
|
|
|
|
last; |
518
|
|
|
|
|
|
|
}; |
519
|
|
|
|
|
|
|
# are they blocked? force-start them if they are |
520
|
0
|
|
|
|
|
|
my $waiting_on_inputs = 0; |
521
|
0
|
|
|
|
|
|
foreach my $i (0..1) { |
522
|
0
|
|
|
|
|
|
my $work_ref = $work_depth_ref->[$i]; |
523
|
0
|
0
|
|
|
|
|
if ($work_ref->[0] == -1) { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
524
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: depth $depth forced start on $work_ref->[1].\n" if ($self->{_debug}); |
525
|
0
|
|
|
|
|
|
&{$work_ref->[2]}($work_ref); # start it |
|
0
|
|
|
|
|
|
|
526
|
0
|
|
|
|
|
|
$waiting_on_inputs++; |
527
|
|
|
|
|
|
|
} elsif ($work_ref->[0] == 0) { |
528
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: depth $depth waiting on working " . _pretty_fn($work_ref->[1]) . ".\n" if ($self->{_debug}); |
529
|
0
|
|
|
|
|
|
$waiting_on_inputs++; |
530
|
|
|
|
|
|
|
} elsif ($work_ref->[0] == 1) { |
531
|
|
|
|
|
|
|
# running, so move fred to zombie queue; otherwise ok |
532
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: pushing job to zombie list.\n" if ($self->{_debug}); |
533
|
0
|
|
|
|
|
|
push(@{$self->{_zombie_work}}, $work_ref); |
|
0
|
|
|
|
|
|
|
534
|
|
|
|
|
|
|
} elsif ($work_ref->[0] == 2) { |
535
|
|
|
|
|
|
|
# input is done |
536
|
|
|
|
|
|
|
} else { |
537
|
0
|
|
|
|
|
|
die "interal error: unknown status $work_ref->[0]\n"; |
538
|
|
|
|
|
|
|
}; |
539
|
|
|
|
|
|
|
}; |
540
|
|
|
|
|
|
|
# bail out if inputs are not done yet. |
541
|
0
|
0
|
|
|
|
|
return if ($waiting_on_inputs); |
542
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
# now we KNOW we do not have blocked work |
544
|
0
|
|
|
|
|
|
my(@two_fn) = (shift @${work_depth_ref}, shift @${work_depth_ref}); |
545
|
0
|
|
|
|
|
|
my $output_type = 'file'; |
546
|
0
|
0
|
0
|
|
|
|
if ($closed && $#{$work_depth_ref} == -1 && $depth == $#{$self->{_work}}) { |
|
0
|
0
|
0
|
|
|
|
|
|
0
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
547
|
0
|
|
|
|
|
|
$output_type = 'final'; |
548
|
|
|
|
|
|
|
} elsif ($self->{_endgame} && $closed && $self->{_work_max_files}[$depth] <= $self->{_endgame_max_files}) { |
549
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: endgame parallelism.\n" if ($self->{_debug}); |
550
|
0
|
|
|
|
|
|
$output_type = 'ipc'; |
551
|
|
|
|
|
|
|
}; |
552
|
0
|
|
|
|
|
|
my($out_fn) = $self->segment_next_output($output_type); |
553
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: depth $depth planning " . _pretty_fn($two_fn[0][1]) . " and " . _pretty_fn($two_fn[1][1]) . " to " . _pretty_fn($out_fn) . ".\n" if ($self->{_debug}); |
554
|
|
|
|
|
|
|
|
555
|
0
|
|
|
|
|
|
foreach my $i (0..1) { |
556
|
0
|
0
|
0
|
|
|
|
next if (ref($two_fn[$i][1]) =~ /^(Fsdb::BoundedQueue|IO::)/ || $two_fn[$i][1] eq '-'); |
557
|
0
|
0
|
|
|
|
|
croak $self->{_prog} . ": file $two_fn[$i][1] is missing.\n" |
558
|
|
|
|
|
|
|
if (! -f $two_fn[$i][1]); |
559
|
|
|
|
|
|
|
}; |
560
|
|
|
|
|
|
|
|
561
|
0
|
0
|
|
|
|
|
if ($output_type eq 'final') { |
562
|
|
|
|
|
|
|
# last time: do it here, in-line |
563
|
|
|
|
|
|
|
# so that we update $self->{_out} in the main thread |
564
|
0
|
|
|
|
|
|
$self->segments_merge2_run($out_fn, 1, $two_fn[0][1], $two_fn[1][1], $self->_unique_id()); |
565
|
0
|
|
|
|
|
|
return; |
566
|
|
|
|
|
|
|
}; |
567
|
|
|
|
|
|
|
# |
568
|
|
|
|
|
|
|
# fork a Fred to do the merge |
569
|
|
|
|
|
|
|
# |
570
|
0
|
0
|
|
|
|
|
my $out_fn_reader = (ref($out_fn) eq 'ARRAY') ? $out_fn->[0] : $out_fn; |
571
|
0
|
0
|
|
|
|
|
my $out_fn_writer = (ref($out_fn) eq 'ARRAY') ? $out_fn->[1] : $out_fn; |
572
|
0
|
|
|
|
|
|
$out_fn = undef; |
573
|
0
|
|
|
|
|
|
my $new_work_ref = [-1, $out_fn_reader, undef]; |
574
|
0
|
|
|
|
|
|
my $unique_id = $self->_unique_id(); |
575
|
0
|
|
|
|
|
|
my $desc = "dbmerge2($two_fn[0][1],$two_fn[1][1]) => $out_fn_reader (id $unique_id)"; |
576
|
|
|
|
|
|
|
my $start_sub = sub { |
577
|
0
|
|
|
0
|
|
|
$self->{_parallelism_available}--; |
578
|
0
|
0
|
|
|
|
|
$new_work_ref->[0] = ($output_type eq 'ipc' ? 1 : 0); # running |
579
|
|
|
|
|
|
|
my $fred = new Fsdb::Support::Freds($desc, sub { |
580
|
0
|
0
|
|
|
|
|
$out_fn_reader->close if ($output_type eq 'ipc'); |
581
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: Fred start $desc\n" if ($self->{_debug}); |
582
|
0
|
|
|
|
|
|
$self->segments_merge2_run($out_fn_writer, 0, $two_fn[0][1], $two_fn[1][1], $unique_id); |
583
|
0
|
0
|
0
|
|
|
|
sleep(1) if (defined($self->{_test}) && $self->{_test} eq 'delay-finish'); |
584
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: Fred end $desc\n" if ($self->{_debug}); |
585
|
0
|
|
|
|
|
|
exit 0; |
586
|
|
|
|
|
|
|
}, sub { |
587
|
0
|
|
|
|
|
|
my($done_fred, $exit_code) = @_; |
588
|
|
|
|
|
|
|
# we're done! |
589
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: Fred post-mortem $desc\n" if ($self->{_debug}); |
590
|
|
|
|
|
|
|
# xxx: with TEST/dbmerge_3_input.cmd I sometimes get exit code 255 (!) although things complete. |
591
|
|
|
|
|
|
|
# turned out Fsdb::Support::Freds::END was messing with $?. |
592
|
0
|
0
|
|
|
|
|
croak "dbmerge: merge2 subprocess $desc, exit code: $exit_code\n" if ($exit_code != 0); |
593
|
0
|
|
|
|
|
|
$new_work_ref->[0] = 2; # done |
594
|
0
|
|
|
|
|
|
}); |
595
|
0
|
|
|
|
|
|
$new_work_ref->[2] = $fred; |
596
|
0
|
0
|
|
|
|
|
$out_fn_writer->close if ($output_type eq 'ipc'); # discard |
597
|
0
|
|
|
|
|
|
}; |
598
|
0
|
|
|
|
|
|
$new_work_ref->[2] = $start_sub; |
599
|
|
|
|
|
|
|
# Put the thread in our queue, |
600
|
|
|
|
|
|
|
# and run it if it's important (pipe or final), |
601
|
|
|
|
|
|
|
# or if we have the capacity. |
602
|
0
|
|
|
|
|
|
$self->enqueue_work($depth + 1, $new_work_ref); |
603
|
0
|
0
|
0
|
|
|
|
&$start_sub() if ($output_type ne 'file' || $self->{_parallelism_available} > 0); |
604
|
0
|
0
|
|
|
|
|
print "# segments_merge_one_depth: looping after $desc.\n" if ($self->{_debug}); |
605
|
|
|
|
|
|
|
}; |
606
|
|
|
|
|
|
|
# At this point all possible work has been queued and maybe started. |
607
|
|
|
|
|
|
|
# If the depth is closed, the work should be empty. |
608
|
|
|
|
|
|
|
# if not, there may be some files in the queue |
609
|
|
|
|
|
|
|
} |
610
|
|
|
|
|
|
|
|
611
|
|
|
|
|
|
|
=head2 segments_xargs |
612
|
|
|
|
|
|
|
|
613
|
|
|
|
|
|
|
$self->segments_xargs(); |
614
|
|
|
|
|
|
|
|
615
|
|
|
|
|
|
|
Internal: read new filenames to process (from stdin) |
616
|
|
|
|
|
|
|
and send them to the work queue. |
617
|
|
|
|
|
|
|
|
618
|
|
|
|
|
|
|
Making a separate Fred to handle xargs is a lot of work, |
619
|
|
|
|
|
|
|
but it guarantees it comes in on an IO::Handle that is selectable. |
620
|
|
|
|
|
|
|
|
621
|
|
|
|
|
|
|
=cut |
622
|
|
|
|
|
|
|
|
623
|
|
|
|
|
|
|
sub segments_xargs($) { |
624
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
625
|
0
|
|
|
|
|
|
my $ipc = $self->{_xargs_ipc_writer}; |
626
|
|
|
|
|
|
|
|
627
|
0
|
|
|
|
|
|
my $num_inputs = 0; |
628
|
|
|
|
|
|
|
|
629
|
|
|
|
|
|
|
# read files as in fsdb format |
630
|
0
|
0
|
|
|
|
|
if ($#{$self->{_inputs}} == 0) { |
|
0
|
|
|
|
|
|
|
631
|
|
|
|
|
|
|
# hacky... |
632
|
0
|
|
|
|
|
|
$self->{_input} = $self->{_inputs}[0]; |
633
|
|
|
|
|
|
|
}; |
634
|
0
|
|
|
|
|
|
$self->finish_io_option('input', -header => '#fsdb filename'); |
635
|
0
|
|
|
|
|
|
my $read_fastpath_sub = $self->{_in}->fastpath_sub(); |
636
|
0
|
|
|
|
|
|
while (my $fref = &$read_fastpath_sub()) { |
637
|
|
|
|
|
|
|
# send each file for processing as level zero |
638
|
0
|
0
|
|
|
|
|
print "# dbmerge: segments_xargs: got $fref->[0]\n" if ($self->{_debug}); |
639
|
0
|
|
|
|
|
|
$ipc->print($fref->[0] . "\n"); |
640
|
0
|
|
|
|
|
|
$num_inputs++; |
641
|
|
|
|
|
|
|
}; |
642
|
|
|
|
|
|
|
# if ($num_inputs <= 1) { |
643
|
|
|
|
|
|
|
# $ipc->print("-1\terror: --xargs, but zero or one input files; dbmerge needs at least two.\n"); # signal eof-f |
644
|
|
|
|
|
|
|
# }; |
645
|
0
|
|
|
|
|
|
$ipc->close; |
646
|
0
|
|
|
|
|
|
$self->{_ipc_writer} = undef; |
647
|
|
|
|
|
|
|
} |
648
|
|
|
|
|
|
|
|
649
|
|
|
|
|
|
|
|
650
|
|
|
|
|
|
|
=head2 segments_merge_all |
651
|
|
|
|
|
|
|
|
652
|
|
|
|
|
|
|
$self->segments_merge_all() |
653
|
|
|
|
|
|
|
|
654
|
|
|
|
|
|
|
Internal: |
655
|
|
|
|
|
|
|
Merge queued files, if any. |
656
|
|
|
|
|
|
|
Iterates over all depths of the merge tree, |
657
|
|
|
|
|
|
|
and handles any forked threads. |
658
|
|
|
|
|
|
|
|
659
|
|
|
|
|
|
|
=head3 Merging Strategy |
660
|
|
|
|
|
|
|
|
661
|
|
|
|
|
|
|
Merging is done in a binary tree is managed through the C<_work> queue. |
662
|
|
|
|
|
|
|
It has an array of C entries, |
663
|
|
|
|
|
|
|
one for each level of the tree. |
664
|
|
|
|
|
|
|
|
665
|
|
|
|
|
|
|
Items are processed in order at each level of the tree, |
666
|
|
|
|
|
|
|
and only level-by-level, so the sort is stable. |
667
|
|
|
|
|
|
|
|
668
|
|
|
|
|
|
|
=head3 Parallelism Model |
669
|
|
|
|
|
|
|
|
670
|
|
|
|
|
|
|
Parallelism is also managed through the C<_work> queue, |
671
|
|
|
|
|
|
|
each element of which consists of one file or stream suitable for merging. |
672
|
|
|
|
|
|
|
The work queue contains both ready output (files or BoundedQueue streams) |
673
|
|
|
|
|
|
|
that can be immediately handled, and pairs of semaphore/pending output |
674
|
|
|
|
|
|
|
for work that is not yet started. |
675
|
|
|
|
|
|
|
All manipulation of the work queue happens in the main thread |
676
|
|
|
|
|
|
|
(with C and C). |
677
|
|
|
|
|
|
|
|
678
|
|
|
|
|
|
|
We start a thread to handle each item in the work queue, |
679
|
|
|
|
|
|
|
and limit parallelism to the C<_max_parallelism>, |
680
|
|
|
|
|
|
|
defaulting to the number of available processors. |
681
|
|
|
|
|
|
|
|
682
|
|
|
|
|
|
|
There two two kinds of parallelism, regular and endgame. |
683
|
|
|
|
|
|
|
For regular parallelism we pick two items off the work queue, |
684
|
|
|
|
|
|
|
merge them, and put the result back on the queue as a new file. |
685
|
|
|
|
|
|
|
Items in the work queue may not be ready. For in-progress items we |
686
|
|
|
|
|
|
|
wait until they are done. For not-yet-started items |
687
|
|
|
|
|
|
|
we start them, then wait until they are done. |
688
|
|
|
|
|
|
|
|
689
|
|
|
|
|
|
|
Endgame parallelism handles the final stages of a large merge. |
690
|
|
|
|
|
|
|
When there are enough processors that we can start a merge jobs |
691
|
|
|
|
|
|
|
for all remaining levels of the merge tree. At this point we switch |
692
|
|
|
|
|
|
|
from merging to files to merging into C pipelines |
693
|
|
|
|
|
|
|
that connect merge processes which start and run concurrently. |
694
|
|
|
|
|
|
|
|
695
|
|
|
|
|
|
|
The final merge is done in the main thread so that that the main thread |
696
|
|
|
|
|
|
|
can handle the output stream and recording the merge action. |
697
|
|
|
|
|
|
|
|
698
|
|
|
|
|
|
|
=cut |
699
|
|
|
|
|
|
|
|
700
|
|
|
|
|
|
|
sub segments_merge_all($) { |
701
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
702
|
|
|
|
|
|
|
|
703
|
0
|
|
|
|
|
|
my $xargs_select; |
704
|
0
|
0
|
|
|
|
|
if ($self->{_xargs}) { |
705
|
0
|
|
|
|
|
|
$xargs_select = IO::Select->new(); |
706
|
0
|
|
|
|
|
|
$xargs_select->add($self->{_xargs_ipc_reader}); |
707
|
|
|
|
|
|
|
}; |
708
|
|
|
|
|
|
|
|
709
|
|
|
|
|
|
|
# |
710
|
|
|
|
|
|
|
# Alternate forking off new merges from finished files |
711
|
|
|
|
|
|
|
# and reaping finished processes. |
712
|
|
|
|
|
|
|
# |
713
|
0
|
|
|
|
|
|
my $overall_progress = 0; |
714
|
0
|
|
|
|
|
|
my $PROGRESS_START = 0.001; |
715
|
0
|
|
|
|
|
|
my $PROGRESS_MAX = 2; |
716
|
0
|
|
|
|
|
|
my $PROGRESS_MULTIPLIER = 2; |
717
|
0
|
|
|
|
|
|
my $progress_backoff = $PROGRESS_START; |
718
|
0
|
|
|
|
|
|
for (;;) { |
719
|
|
|
|
|
|
|
# done? |
720
|
0
|
|
|
|
|
|
my $deepest = $#{$self->{_work}}; |
|
0
|
|
|
|
|
|
|
721
|
0
|
0
|
0
|
|
|
|
last if ($self->{_work_closed}[$deepest] && $#{$self->{_work}[$deepest]} <= 0); |
|
0
|
|
|
|
|
|
|
722
|
|
|
|
|
|
|
|
723
|
|
|
|
|
|
|
# |
724
|
|
|
|
|
|
|
# First, put more work on the queue, where possible. |
725
|
|
|
|
|
|
|
# |
726
|
|
|
|
|
|
|
# Go through this loop multiple times because closing the last depth |
727
|
|
|
|
|
|
|
# can actually allow work to start at the last+1 depth, |
728
|
|
|
|
|
|
|
# and in endgame mode we risk blocking (due to flow control) |
729
|
|
|
|
|
|
|
# if we don't start threads at all depths. |
730
|
|
|
|
|
|
|
# |
731
|
0
|
|
|
|
|
|
my $try_again = 1; |
732
|
0
|
|
|
|
|
|
while ($try_again) { |
733
|
0
|
|
|
|
|
|
foreach my $depth (0..$#{$self->{_work}}) { |
|
0
|
|
|
|
|
|
|
734
|
0
|
|
|
|
|
|
$self->segments_merge_one_depth($depth); |
735
|
0
|
|
|
|
|
|
$try_again = undef; |
736
|
0
|
0
|
0
|
|
|
|
if ($#{$self->{_work}[$depth]} == -1 && $self->{_work_closed}[$depth]) { |
|
0
|
|
|
|
|
|
|
737
|
|
|
|
|
|
|
# When one level is all in progress, we can close the next. |
738
|
0
|
|
|
|
|
|
my $next_depth = $depth + 1; |
739
|
0
|
0
|
|
|
|
|
if (!$self->{_work_closed}[$next_depth]) { |
740
|
0
|
|
|
|
|
|
$self->{_work_closed}[$next_depth] = 1; |
741
|
0
|
|
|
|
|
|
$try_again = 1; |
742
|
0
|
|
|
|
|
|
$overall_progress++; |
743
|
0
|
0
|
|
|
|
|
print "# segments_merge_all: closed work depth $next_depth\n" if ($self->{_debug}); |
744
|
|
|
|
|
|
|
}; |
745
|
|
|
|
|
|
|
}; |
746
|
|
|
|
|
|
|
}; |
747
|
|
|
|
|
|
|
}; |
748
|
|
|
|
|
|
|
|
749
|
|
|
|
|
|
|
# |
750
|
|
|
|
|
|
|
# Next, handle Freds that have finished. |
751
|
|
|
|
|
|
|
# Reap as many as possible. |
752
|
|
|
|
|
|
|
# |
753
|
0
|
0
|
|
|
|
|
print "# segments_merge_all: reaping threads\n" if ($self->{_debug}); |
754
|
0
|
|
|
|
|
|
for (;;) { |
755
|
0
|
|
|
|
|
|
my $fred_or_code = Fsdb::Support::Freds::join_any(); |
756
|
0
|
0
|
|
|
|
|
last if (ref($fred_or_code) eq ''); |
757
|
0
|
|
|
|
|
|
$overall_progress++; |
758
|
0
|
0
|
|
|
|
|
croak "dbmerge: merge thread failed\n" |
759
|
|
|
|
|
|
|
if ($fred_or_code->exit_code() != 0); |
760
|
0
|
0
|
|
|
|
|
print "# segments_merge_all: merged fred " . $fred_or_code->info() . "\n" if ($self->{_debug}); |
761
|
|
|
|
|
|
|
}; |
762
|
|
|
|
|
|
|
|
763
|
|
|
|
|
|
|
# |
764
|
|
|
|
|
|
|
# Now start up more parallelism, if possible. |
765
|
|
|
|
|
|
|
# |
766
|
0
|
|
|
|
|
|
my $depth = 0; |
767
|
0
|
|
|
|
|
|
my $i = 0; |
768
|
0
|
|
|
|
|
|
while ($self->{_parallelism_available} > 0) { |
769
|
0
|
|
|
|
|
|
my $work_ref = $self->{_work}[$depth][$i]; |
770
|
0
|
0
|
0
|
|
|
|
if (defined($work_ref) && $work_ref->[0] == -1) { |
771
|
|
|
|
|
|
|
# start it (it will decrement parallelism) |
772
|
0
|
|
|
|
|
|
&{$work_ref->[2]}($work_ref); |
|
0
|
|
|
|
|
|
|
773
|
0
|
|
|
|
|
|
$overall_progress++; |
774
|
|
|
|
|
|
|
}; |
775
|
|
|
|
|
|
|
# walk the whole work queue |
776
|
0
|
0
|
|
|
|
|
if (++$i > $#{$self->{_work}[$depth]}) { |
|
0
|
|
|
|
|
|
|
777
|
0
|
0
|
|
|
|
|
last if (++$depth > $#{$self->{_work}}); |
|
0
|
|
|
|
|
|
|
778
|
0
|
|
|
|
|
|
$i = 0; |
779
|
|
|
|
|
|
|
}; |
780
|
|
|
|
|
|
|
}; |
781
|
|
|
|
|
|
|
|
782
|
|
|
|
|
|
|
# |
783
|
|
|
|
|
|
|
# Handle xargs, if any. |
784
|
|
|
|
|
|
|
# |
785
|
|
|
|
|
|
|
# Unfortunately, we busy-loop here, |
786
|
|
|
|
|
|
|
# because we need to alternate reaping finished processes |
787
|
|
|
|
|
|
|
# and xargs. |
788
|
|
|
|
|
|
|
# |
789
|
|
|
|
|
|
|
# Fortunately, this terminates when xargs are complete. |
790
|
|
|
|
|
|
|
# |
791
|
0
|
0
|
|
|
|
|
if ($self->{_xargs_ipc_status}) { |
792
|
0
|
|
|
|
|
|
my(@ready) = $xargs_select->can_read($progress_backoff); |
793
|
0
|
|
|
|
|
|
foreach my $fh (@ready) { |
794
|
0
|
|
|
|
|
|
my ($fn) = $fh->getline; |
795
|
0
|
0
|
|
|
|
|
if (defined($fn)) { |
796
|
0
|
|
|
|
|
|
chomp($fn); |
797
|
|
|
|
|
|
|
$self->{_files_cleanup}{$fn} = 'unlink' |
798
|
0
|
0
|
|
|
|
|
if ($self->{_remove_inputs}); |
799
|
0
|
|
|
|
|
|
$self->enqueue_work(0, [2, $fn, undef]); |
800
|
0
|
0
|
|
|
|
|
print "# xargs receive $fn\n" if ($self->{_debug}); |
801
|
|
|
|
|
|
|
} else { |
802
|
|
|
|
|
|
|
# eof, so no more xargs |
803
|
0
|
|
|
|
|
|
$self->{_work_closed}[0] = 1; |
804
|
0
|
|
|
|
|
|
$self->{_xargs_ipc_status} = undef; |
805
|
|
|
|
|
|
|
}; |
806
|
0
|
|
|
|
|
|
$overall_progress++; |
807
|
|
|
|
|
|
|
}; |
808
|
|
|
|
|
|
|
}; |
809
|
|
|
|
|
|
|
# |
810
|
|
|
|
|
|
|
# Avoid spinlooping. |
811
|
|
|
|
|
|
|
# |
812
|
0
|
0
|
|
|
|
|
if (!$overall_progress) { |
813
|
|
|
|
|
|
|
# No progress, so stall. |
814
|
0
|
0
|
|
|
|
|
print "# segments_merge_all: stalling for $progress_backoff\n" if ($self->{_debug}); |
815
|
0
|
|
|
|
|
|
sleep($progress_backoff); |
816
|
0
|
|
|
|
|
|
$progress_backoff *= $PROGRESS_MULTIPLIER; # exponential backoff |
817
|
0
|
0
|
|
|
|
|
$progress_backoff = $PROGRESS_MAX if ($progress_backoff > $PROGRESS_MAX); |
818
|
|
|
|
|
|
|
} else { |
819
|
|
|
|
|
|
|
# Woo-hoo, did something. Rush right back and try again. |
820
|
0
|
|
|
|
|
|
$overall_progress = 0; |
821
|
0
|
|
|
|
|
|
$progress_backoff = $PROGRESS_START; |
822
|
|
|
|
|
|
|
}; |
823
|
|
|
|
|
|
|
}; |
824
|
|
|
|
|
|
|
|
825
|
|
|
|
|
|
|
# reap endgame zombies |
826
|
0
|
|
|
|
|
|
while (my $zombie_work_ref = shift(@{$self->{_zombie_work}})) { |
|
0
|
|
|
|
|
|
|
827
|
0
|
0
|
|
|
|
|
next if ($zombie_work_ref->[0] == 2); |
828
|
0
|
0
|
|
|
|
|
print "# waiting on zombie " . $zombie_work_ref->[2]->info() . "\n" if ($self->{_debug}); |
829
|
0
|
|
|
|
|
|
$zombie_work_ref->[2]->join(); |
830
|
0
|
0
|
|
|
|
|
croak "internal error: zombie didn't reset status\n" if ($zombie_work_ref->[0] != 2); |
831
|
|
|
|
|
|
|
}; |
832
|
|
|
|
|
|
|
|
833
|
|
|
|
|
|
|
# reap xargs (if it didn't already get picked up) |
834
|
0
|
0
|
|
|
|
|
if ($self->{_xargs_fred}) { |
835
|
0
|
0
|
|
|
|
|
print "# waiting on xargs fred\n" if ($self->{_debug}); |
836
|
0
|
|
|
|
|
|
$self->{_xargs_fred}->join(); |
837
|
|
|
|
|
|
|
}; |
838
|
|
|
|
|
|
|
# |
839
|
|
|
|
|
|
|
} |
840
|
|
|
|
|
|
|
|
841
|
|
|
|
|
|
|
|
842
|
|
|
|
|
|
|
|
843
|
|
|
|
|
|
|
=head2 setup |
844
|
|
|
|
|
|
|
|
845
|
|
|
|
|
|
|
$filter->setup(); |
846
|
|
|
|
|
|
|
|
847
|
|
|
|
|
|
|
Internal: setup, parse headers. |
848
|
|
|
|
|
|
|
|
849
|
|
|
|
|
|
|
=cut |
850
|
|
|
|
|
|
|
|
851
|
|
|
|
|
|
|
sub setup($) { |
852
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
853
|
|
|
|
|
|
|
|
854
|
|
|
|
|
|
|
croak $self->{_prog} . ": no sorting key specified.\n" |
855
|
0
|
0
|
|
|
|
|
if ($#{$self->{_sort_argv}} == -1); |
|
0
|
|
|
|
|
|
|
856
|
|
|
|
|
|
|
|
857
|
0
|
0
|
0
|
|
|
|
if (!$self->{_xargs} && $#{$self->{_inputs}} == -1) { |
|
0
|
|
|
|
|
|
|
858
|
0
|
|
|
|
|
|
croak $self->{_prog} . ": no input sources specified, use --input or --xargs.\n"; |
859
|
|
|
|
|
|
|
}; |
860
|
0
|
0
|
0
|
|
|
|
if (!$self->{_xargs} && $#{$self->{_inputs}} == 0) { |
|
0
|
|
|
|
|
|
|
861
|
0
|
|
|
|
|
|
croak $self->{_prog} . ": only one input source, but can't merge one file.\n"; |
862
|
|
|
|
|
|
|
}; |
863
|
0
|
0
|
0
|
|
|
|
if ($self->{_xargs} && $#{$self->{_inputs}} > 0) { |
|
0
|
|
|
|
|
|
|
864
|
0
|
|
|
|
|
|
croak $self->{_prog} . ": --xargs and multiple inputs (perhaps you meant NOT --xargs?).\n"; |
865
|
|
|
|
|
|
|
}; |
866
|
|
|
|
|
|
|
# prove files exist (early error checking) |
867
|
0
|
|
|
|
|
|
foreach (@{$self->{_inputs}}) { |
|
0
|
|
|
|
|
|
|
868
|
0
|
0
|
|
|
|
|
next if (ref($_) ne ''); # skip objects |
869
|
0
|
0
|
|
|
|
|
next if ($_ eq '-'); # special case: stdin |
870
|
0
|
0
|
|
|
|
|
if (! -f $_) { |
871
|
0
|
|
|
|
|
|
croak $self->{_prog} . ": input source $_ does not exist.\n"; |
872
|
|
|
|
|
|
|
}; |
873
|
|
|
|
|
|
|
}; |
874
|
0
|
0
|
|
|
|
|
if ($self->{_remove_inputs}) { |
875
|
0
|
|
|
|
|
|
foreach (@{$self->{_inputs}}) { |
|
0
|
|
|
|
|
|
|
876
|
0
|
0
|
|
|
|
|
$self->{_files_cleanup}{$_} = 'unlink' |
877
|
|
|
|
|
|
|
if ($_ ne '-'); |
878
|
|
|
|
|
|
|
}; |
879
|
|
|
|
|
|
|
}; |
880
|
|
|
|
|
|
|
# |
881
|
|
|
|
|
|
|
# the _work queue consists of |
882
|
|
|
|
|
|
|
# 1. [2, filenames, fred] that for completed files need to be merged. |
883
|
|
|
|
|
|
|
# 1a. [1, IO::Handle::Pipe, fred] for files that are being processed |
884
|
|
|
|
|
|
|
# and can be merged (but are not yet done). |
885
|
|
|
|
|
|
|
# 2. [-1, filename, $start_sub] for blocked threads that, when started |
886
|
|
|
|
|
|
|
# by evaling $start_sub, will go to filename. |
887
|
|
|
|
|
|
|
# 3. [0, filename, fred] for files in the process of being computed |
888
|
|
|
|
|
|
|
# |
889
|
|
|
|
|
|
|
# Filename can be an Fsdb::BoundedQueue or IO::Pipe objects for endgame mode threads |
890
|
|
|
|
|
|
|
# |
891
|
|
|
|
|
|
|
# _work_closed is set when that depth is no longer growing; |
892
|
|
|
|
|
|
|
# at that time _work_depth_files is the maximum number of files there. |
893
|
|
|
|
|
|
|
# |
894
|
|
|
|
|
|
|
# Put stuff on it with $self->enqueue_work to keep the |
895
|
|
|
|
|
|
|
# related variables correct. |
896
|
|
|
|
|
|
|
# |
897
|
0
|
|
|
|
|
|
$self->{_work}[0] = []; |
898
|
0
|
|
|
|
|
|
$self->{_work_closed}[0] = 0; |
899
|
0
|
|
|
|
|
|
$self->{_work_depth_files}[0] = 0; |
900
|
0
|
|
|
|
|
|
$self->{_xargs_ipc_status} = undef; |
901
|
0
|
|
|
|
|
|
$self->{_zombie_work} = []; # collect in-process but not-yet-done freds |
902
|
0
|
0
|
|
|
|
|
if (!$self->{_xargs}) { |
903
|
0
|
|
|
|
|
|
foreach (@{$self->{_inputs}}) { |
|
0
|
|
|
|
|
|
|
904
|
0
|
|
|
|
|
|
$self->enqueue_work(0, [2, $_, undef]); |
905
|
|
|
|
|
|
|
}; |
906
|
0
|
|
|
|
|
|
$self->{_work_closed}[0] = 1; |
907
|
|
|
|
|
|
|
} else { |
908
|
0
|
|
|
|
|
|
my $xargs_ipc_reader = new IO::Handle; |
909
|
0
|
|
|
|
|
|
my $xargs_ipc_writer = new IO::Handle; |
910
|
0
|
0
|
|
|
|
|
pipe($xargs_ipc_reader, $xargs_ipc_writer) or croak "cannot open pipe\n"; |
911
|
0
|
|
|
|
|
|
$self->{_xargs_ipc_status} = 'running'; |
912
|
|
|
|
|
|
|
$self->{_xargs_fred} = new Fsdb::Support::Freds('dbmerge:xargs', |
913
|
|
|
|
|
|
|
sub { |
914
|
0
|
|
|
0
|
|
|
$SIG{PIPE} = 'IGNORE'; # in case we finish before main reads anything |
915
|
0
|
|
|
|
|
|
$xargs_ipc_reader->close; |
916
|
0
|
|
|
|
|
|
$xargs_ipc_writer->autoflush(1); |
917
|
0
|
|
|
|
|
|
$self->{_xargs_ipc_writer} = $xargs_ipc_writer; |
918
|
0
|
|
|
|
|
|
$self->segments_xargs(); |
919
|
0
|
|
|
|
|
|
exit 0; |
920
|
|
|
|
|
|
|
}, sub { |
921
|
0
|
|
|
0
|
|
|
$self->{_xargs_ipc_status} = 'completed'; |
922
|
0
|
|
|
|
|
|
}); |
923
|
0
|
|
|
|
|
|
$xargs_ipc_reader->autoflush(1); |
924
|
0
|
|
|
|
|
|
$xargs_ipc_writer->close; |
925
|
0
|
|
|
|
|
|
$self->{_xargs_ipc_reader} = $xargs_ipc_reader; |
926
|
|
|
|
|
|
|
# actual xargs reception happens in our main loop in segments_merge_all() |
927
|
|
|
|
|
|
|
}; |
928
|
|
|
|
|
|
|
# |
929
|
|
|
|
|
|
|
# control parallelism |
930
|
|
|
|
|
|
|
# |
931
|
|
|
|
|
|
|
# For the endgame, we overcommit by a large factor |
932
|
|
|
|
|
|
|
# because in the merge tree many become blocked on the IO pipeline. |
933
|
|
|
|
|
|
|
# |
934
|
0
|
|
0
|
|
|
|
$self->{_max_parallelism} //= Fsdb::Support::OS::max_parallelism(); |
935
|
0
|
|
0
|
|
|
|
$self->{_parallelism_available} //= $self->{_max_parallelism}; |
936
|
0
|
|
|
|
|
|
my $viable_endgame_processes = $self->{_max_parallelism}; |
937
|
0
|
|
|
|
|
|
my $files_to_merge = 1; |
938
|
0
|
|
|
|
|
|
while ($viable_endgame_processes > 0) { |
939
|
0
|
|
|
|
|
|
$viable_endgame_processes -= $files_to_merge; |
940
|
0
|
|
|
|
|
|
$files_to_merge *= 2; |
941
|
|
|
|
|
|
|
}; |
942
|
0
|
|
|
|
|
|
$self->{_endgame_max_files} = int($files_to_merge); |
943
|
0
|
0
|
|
|
|
|
STDOUT->autoflush(1) if ($self->{_debug}); |
944
|
0
|
0
|
|
|
|
|
print "# dbmerge: endgame_max_files: " . $self->{_endgame_max_files} . "\n" if($self->{_debug}); |
945
|
|
|
|
|
|
|
} |
946
|
|
|
|
|
|
|
|
947
|
|
|
|
|
|
|
=head2 run |
948
|
|
|
|
|
|
|
|
949
|
|
|
|
|
|
|
$filter->run(); |
950
|
|
|
|
|
|
|
|
951
|
|
|
|
|
|
|
Internal: run over each rows. |
952
|
|
|
|
|
|
|
|
953
|
|
|
|
|
|
|
=cut |
954
|
|
|
|
|
|
|
sub run($) { |
955
|
0
|
|
|
0
|
1
|
|
my($self) = @_; |
956
|
|
|
|
|
|
|
|
957
|
0
|
|
|
|
|
|
$self->segments_merge_all(); |
958
|
|
|
|
|
|
|
}; |
959
|
|
|
|
|
|
|
|
960
|
|
|
|
|
|
|
|
961
|
|
|
|
|
|
|
|
962
|
|
|
|
|
|
|
=head1 AUTHOR and COPYRIGHT |
963
|
|
|
|
|
|
|
|
964
|
|
|
|
|
|
|
Copyright (C) 1991-2015 by John Heidemann |
965
|
|
|
|
|
|
|
|
966
|
|
|
|
|
|
|
This program is distributed under terms of the GNU general |
967
|
|
|
|
|
|
|
public license, version 2. See the file COPYING |
968
|
|
|
|
|
|
|
with the distribution for details. |
969
|
|
|
|
|
|
|
|
970
|
|
|
|
|
|
|
=cut |
971
|
|
|
|
|
|
|
|
972
|
|
|
|
|
|
|
1; |