@recaptime-dev's working patches + fork for Phorge, a community fork of Phabricator. (Upstream dev and stable branches are at upstream/main and upstream/stable respectively.)
hq.recaptime.dev/wiki/Phorge
phorge
phabricator
Inform about moved code and prefer it over copied code
Summary:
Also reduce the memory usage a little bit (before increasing it again).
I use the same CSS class as for the copied code.
Test Plan: Parsed 100 diffs and checked about 10 of them - looks good.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Koolvin
Differential Revision: https://secure.phabricator.com/D2339
+104
-42
+48
scripts/differential/detect_copied_code.php
+48
scripts/differential/detect_copied_code.php
···
1
+
#!/usr/bin/env php
2
+
<?php
3
+
4
+
/*
5
+
* Copyright 2012 Facebook, Inc.
6
+
*
7
+
* Licensed under the Apache License, Version 2.0 (the "License");
8
+
* you may not use this file except in compliance with the License.
9
+
* You may obtain a copy of the License at
10
+
*
11
+
* http://www.apache.org/licenses/LICENSE-2.0
12
+
*
13
+
* Unless required by applicable law or agreed to in writing, software
14
+
* distributed under the License is distributed on an "AS IS" BASIS,
15
+
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16
+
* See the License for the specific language governing permissions and
17
+
* limitations under the License.
18
+
*/
19
+
20
+
$root = dirname(dirname(dirname(__FILE__)));
21
+
require_once $root.'/scripts/__init_script__.php';
22
+
23
+
if ($argc != 3 || !is_numeric($argv[1]) || !is_numeric($argv[2])) {
24
+
echo "Usage: {$argv[0]} <diff_id_from> <diff_id_to>\n";
25
+
exit(1);
26
+
}
27
+
list(, $from, $to) = $argv;
28
+
29
+
for ($diff_id = $from; $diff_id <= $to; $diff_id++) {
30
+
echo "Processing $diff_id";
31
+
$diff = id(new DifferentialDiff())->load($diff_id);
32
+
$diff->attachChangesets($diff->loadChangesets());
33
+
$orig_copy = array();
34
+
foreach ($diff->getChangesets() as $i => $changeset) {
35
+
$orig_copy[$i] = idx((array)$changeset->getMetadata(), 'copy:lines');
36
+
$changeset->attachHunks($changeset->loadHunks());
37
+
}
38
+
$diff->detectCopiedCode();
39
+
foreach ($diff->getChangesets() as $i => $changeset) {
40
+
if (idx($changeset->getMetadata(), 'copy:lines') || $orig_copy[$i]) {
41
+
echo ".";
42
+
$changeset->save();
43
+
}
44
+
}
45
+
echo "\n";
46
+
}
47
+
48
+
echo "Done.\n";
+16
-16
src/__celerity_resource_map__.php
+16
-16
src/__celerity_resource_map__.php
···
531
531
),
532
532
'differential-changeset-view-css' =>
533
533
array(
534
-
'uri' => '/res/5c7db62d/rsrc/css/application/differential/changeset-view.css',
534
+
'uri' => '/res/c7edd492/rsrc/css/application/differential/changeset-view.css',
535
535
'type' => 'css',
536
536
'requires' =>
537
537
array(
···
2491
2491
'uri' => '/res/pkg/0c96375e/core.pkg.js',
2492
2492
'type' => 'js',
2493
2493
),
2494
-
'1715d060' =>
2494
+
'acdf073e' =>
2495
2495
array(
2496
2496
'name' => 'differential.pkg.css',
2497
2497
'symbols' =>
···
2510
2510
11 => 'differential-local-commits-view-css',
2511
2511
12 => 'inline-comment-summary-css',
2512
2512
),
2513
-
'uri' => '/res/pkg/1715d060/differential.pkg.css',
2513
+
'uri' => '/res/pkg/acdf073e/differential.pkg.css',
2514
2514
'type' => 'css',
2515
2515
),
2516
2516
70509835 =>
···
2633
2633
'aphront-dialog-view-css' => '9c4e265b',
2634
2634
'aphront-error-view-css' => '9c4e265b',
2635
2635
'aphront-form-view-css' => '9c4e265b',
2636
-
'aphront-headsup-action-list-view-css' => '1715d060',
2636
+
'aphront-headsup-action-list-view-css' => 'acdf073e',
2637
2637
'aphront-headsup-view-css' => '9c4e265b',
2638
2638
'aphront-list-filter-view-css' => '9c4e265b',
2639
2639
'aphront-pager-view-css' => '9c4e265b',
···
2643
2643
'aphront-tokenizer-control-css' => '9c4e265b',
2644
2644
'aphront-tooltip-css' => '9c4e265b',
2645
2645
'aphront-typeahead-control-css' => '9c4e265b',
2646
-
'differential-changeset-view-css' => '1715d060',
2647
-
'differential-core-view-css' => '1715d060',
2646
+
'differential-changeset-view-css' => 'acdf073e',
2647
+
'differential-core-view-css' => 'acdf073e',
2648
2648
'differential-inline-comment-editor' => '70509835',
2649
-
'differential-local-commits-view-css' => '1715d060',
2650
-
'differential-revision-add-comment-css' => '1715d060',
2651
-
'differential-revision-comment-css' => '1715d060',
2652
-
'differential-revision-comment-list-css' => '1715d060',
2653
-
'differential-revision-detail-css' => '1715d060',
2654
-
'differential-revision-history-css' => '1715d060',
2655
-
'differential-table-of-contents-css' => '1715d060',
2649
+
'differential-local-commits-view-css' => 'acdf073e',
2650
+
'differential-revision-add-comment-css' => 'acdf073e',
2651
+
'differential-revision-comment-css' => 'acdf073e',
2652
+
'differential-revision-comment-list-css' => 'acdf073e',
2653
+
'differential-revision-detail-css' => 'acdf073e',
2654
+
'differential-revision-history-css' => 'acdf073e',
2655
+
'differential-table-of-contents-css' => 'acdf073e',
2656
2656
'diffusion-commit-view-css' => 'c8ce2d88',
2657
2657
'diffusion-icons-css' => 'c8ce2d88',
2658
-
'inline-comment-summary-css' => '1715d060',
2658
+
'inline-comment-summary-css' => 'acdf073e',
2659
2659
'javelin-behavior' => '8a5de8a3',
2660
2660
'javelin-behavior-aphront-basic-tokenizer' => '97f65640',
2661
2661
'javelin-behavior-aphront-drag-and-drop' => '70509835',
···
2709
2709
'maniphest-task-summary-css' => '7839ae2d',
2710
2710
'maniphest-transaction-detail-css' => '7839ae2d',
2711
2711
'phabricator-app-buttons-css' => '9c4e265b',
2712
-
'phabricator-content-source-view-css' => '1715d060',
2712
+
'phabricator-content-source-view-css' => 'acdf073e',
2713
2713
'phabricator-core-buttons-css' => '9c4e265b',
2714
2714
'phabricator-core-css' => '9c4e265b',
2715
2715
'phabricator-directory-css' => '9c4e265b',
···
2720
2720
'phabricator-keyboard-shortcut' => '0c96375e',
2721
2721
'phabricator-keyboard-shortcut-manager' => '0c96375e',
2722
2722
'phabricator-menu-item' => '0c96375e',
2723
-
'phabricator-object-selector-css' => '1715d060',
2723
+
'phabricator-object-selector-css' => 'acdf073e',
2724
2724
'phabricator-paste-file-upload' => '0c96375e',
2725
2725
'phabricator-prefab' => '0c96375e',
2726
2726
'phabricator-project-tag-css' => '7839ae2d',
+6
-5
src/applications/differential/parser/changeset/DifferentialChangesetParser.php
+6
-5
src/applications/differential/parser/changeset/DifferentialChangesetParser.php
···
1410
1410
$n_text = $this->new[$ii]['text'];
1411
1411
$n_attr = ' class="comment"';
1412
1412
} else if (isset($copy_lines[$n_num])) {
1413
-
list($orig_file, $orig_line) = $copy_lines[$n_num];
1413
+
list($orig_file, $orig_line, $orig_type) = $copy_lines[$n_num];
1414
+
$title = ($orig_type == '-' ? 'Moved' : 'Copied').' from ';
1414
1415
if ($orig_file == '') {
1415
-
$title = "line {$orig_line}";
1416
+
$title .= "line {$orig_line}";
1416
1417
} else {
1417
-
$title =
1418
+
$title .=
1418
1419
basename($orig_file).
1419
1420
":{$orig_line} in dir ".
1420
1421
dirname('/'.$orig_file);
1421
1422
}
1423
+
$class = ($orig_type == '-' ? 'new-move' : 'new-copy');
1422
1424
$n_attr =
1423
-
' class="new new-copy"'.
1424
-
' title="Copied from '.phutil_escape_html($title).'"';
1425
+
' class="'.$class.'" title="'.phutil_escape_html($title).'"';
1425
1426
} else if (empty($this->old[$ii])) {
1426
1427
$n_attr = ' class="new new-full"';
1427
1428
} else {
+30
-21
src/applications/differential/storage/diff/DifferentialDiff.php
+30
-21
src/applications/differential/storage/diff/DifferentialDiff.php
···
156
156
return $diff;
157
157
}
158
158
159
-
private function detectCopiedCode($min_width = 40, $min_lines = 3) {
159
+
public function detectCopiedCode($min_width = 40, $min_lines = 3) {
160
160
$map = array();
161
161
$files = array();
162
+
$types = array();
162
163
foreach ($this->changesets as $changeset) {
163
164
$file = $changeset->getFilename();
164
165
foreach ($changeset->getHunks() as $hunk) {
165
166
$line = $hunk->getOldOffset();
166
-
foreach (explode("\n", $hunk->makeOldFile()) as $code) {
167
-
$files[$file][$line] = $code;
168
-
if (strlen($code) >= $min_width) {
169
-
$map[$code][] = array($file, $line);
167
+
foreach (explode("\n", $hunk->getChanges()) as $code) {
168
+
$type = (isset($code[0]) ? $code[0] : '');
169
+
if ($type == '-' || $type == ' ') {
170
+
$code = (string)substr($code, 1);
171
+
$files[$file][$line] = $code;
172
+
$types[$file][$line] = $type;
173
+
if (strlen($code) >= $min_width) {
174
+
$map[$code][] = array($file, $line);
175
+
}
176
+
$line++;
170
177
}
171
-
$line++;
172
178
}
173
179
}
174
180
}
···
179
185
$added = $hunk->getAddedLines();
180
186
for (reset($added); list($line, $code) = each($added); next($added)) {
181
187
if (isset($map[$code])) { // We found a long matching line.
182
-
$lengths = array();
183
-
$max_offsets = array();
188
+
$best_length = 0;
184
189
foreach ($map[$code] as $val) { // Explore all candidates.
185
190
list($file, $orig_line) = $val;
186
-
$lengths["$orig_line:$file"] = 1;
191
+
$length = 1;
187
192
// Search also backwards for short lines.
188
193
foreach (array(-1, 1) as $direction) {
189
194
$offset = $direction;
···
191
196
isset($added[$line + $offset]) &&
192
197
idx($files[$file], $orig_line + $offset) ===
193
198
$added[$line + $offset]) {
194
-
$lengths["$orig_line:$file"]++;
199
+
$length++;
195
200
$offset += $direction;
196
201
}
197
202
}
198
-
// ($offset - 1) contains number of forward matching lines.
199
-
$max_offsets["$orig_line:$file"] = $offset - 1;
203
+
if ($length > $best_length ||
204
+
($length == $best_length && // Prefer moves.
205
+
idx($types[$file], $orig_line) == '-')) {
206
+
$best_length = $length;
207
+
// ($offset - 1) contains number of forward matching lines.
208
+
$best_offset = $offset - 1;
209
+
$best_file = $file;
210
+
$best_line = $orig_line;
211
+
}
200
212
}
201
-
$length = max($lengths); // Choose longest candidate.
202
-
$val = array_search($length, $lengths);
203
-
$offset = $max_offsets[$val];
204
-
list($orig_line, $file) = explode(':', $val, 2);
205
-
$save_file = ($file == $changeset->getFilename() ? '' : $file);
206
-
for ($i = $length; $i--; ) {
207
-
$copies[$line + $offset - $i] = ($length < $min_lines
213
+
$file = ($best_file == $changeset->getFilename() ? '' : $best_file);
214
+
for ($i = $best_length; $i--; ) {
215
+
$type = idx($types[$best_file], $best_line + $best_offset - $i);
216
+
$copies[$line + $best_offset - $i] = ($best_length < $min_lines
208
217
? array() // Ignore short blocks.
209
-
: array($save_file, $orig_line + $offset - $i));
218
+
: array($file, $best_line + $best_offset - $i, $type));
210
219
}
211
-
for ($i = 0; $i < $offset; $i++) {
220
+
for ($i = 0; $i < $best_offset; $i++) {
212
221
next($added);
213
222
}
214
223
}