1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
namespace SilverStripe\View\Parsers; |
4
|
|
|
|
5
|
|
|
use InvalidArgumentException; |
6
|
|
|
use SilverStripe\Core\Convert; |
7
|
|
|
use SilverStripe\Core\Injector\Injector; |
8
|
|
|
|
9
|
|
|
require_once 'difflib/difflib.php'; |
10
|
|
|
|
11
|
|
|
/** |
12
|
|
|
* Class representing a 'diff' between two sequences of strings. |
13
|
|
|
*/ |
14
|
|
|
class Diff extends \Diff |
15
|
|
|
{ |
16
|
|
|
public static $html_cleaner_class = null; |
17
|
|
|
|
18
|
|
|
/** |
19
|
|
|
* Attempt to clean invalid HTML, which messes up diffs. |
20
|
|
|
* This cleans code if possible, using an instance of HTMLCleaner |
21
|
|
|
* |
22
|
|
|
* NB: By default, only extremely simple tidying is performed, |
23
|
|
|
* by passing through DomDocument::loadHTML and saveXML |
24
|
|
|
* |
25
|
|
|
* @param string $content HTML content |
26
|
|
|
* @param HTMLCleaner $cleaner Optional instance of a HTMLCleaner class to |
27
|
|
|
* use, overriding self::$html_cleaner_class |
28
|
|
|
* @return mixed|string |
29
|
|
|
*/ |
30
|
|
|
public static function cleanHTML($content, $cleaner = null) |
31
|
|
|
{ |
32
|
|
|
if (!$cleaner) { |
33
|
|
|
if (self::$html_cleaner_class && class_exists(self::$html_cleaner_class)) { |
34
|
|
|
$cleaner = Injector::inst()->create(self::$html_cleaner_class); |
35
|
|
|
} else { |
36
|
|
|
//load cleaner if the dependent class is available |
37
|
|
|
$cleaner = HTMLCleaner::inst(); |
38
|
|
|
} |
39
|
|
|
} |
40
|
|
|
|
41
|
|
|
if ($cleaner) { |
42
|
|
|
$content = $cleaner->cleanHTML($content); |
43
|
|
|
} else { |
44
|
|
|
// At most basic level of cleaning, use DOMDocument to save valid XML. |
45
|
|
|
$doc = HTMLValue::create($content); |
46
|
|
|
$content = $doc->getContent(); |
47
|
|
|
} |
48
|
|
|
|
49
|
|
|
// Remove empty <ins /> and <del /> tags because browsers hate them |
50
|
|
|
$content = preg_replace('/<(ins|del)[^>]*\/>/', '', $content); |
51
|
|
|
|
52
|
|
|
return $content; |
53
|
|
|
} |
54
|
|
|
|
55
|
|
|
/** |
56
|
|
|
* @param string $from |
57
|
|
|
* @param string $to |
58
|
|
|
* @param bool $escape |
59
|
|
|
* @return string |
60
|
|
|
*/ |
61
|
|
|
public static function compareHTML($from, $to, $escape = false) |
62
|
|
|
{ |
63
|
|
|
// First split up the content into words and tags |
64
|
|
|
$set1 = self::getHTMLChunks($from); |
65
|
|
|
$set2 = self::getHTMLChunks($to); |
66
|
|
|
|
67
|
|
|
// Diff that |
68
|
|
|
$diff = new Diff($set1, $set2); |
69
|
|
|
|
70
|
|
|
$tagStack[1] = $tagStack[2] = 0; |
|
|
|
|
71
|
|
|
$rechunked[1] = $rechunked[2] = array(); |
|
|
|
|
72
|
|
|
|
73
|
|
|
// Go through everything, converting edited tags (and their content) into single chunks. Otherwise |
74
|
|
|
// the generated HTML gets crusty |
75
|
|
|
foreach ($diff->edits as $edit) { |
76
|
|
|
$lookForTag = false; |
77
|
|
|
$stuffFor = []; |
78
|
|
|
switch ($edit->type) { |
79
|
|
|
case 'copy': |
80
|
|
|
$lookForTag = false; |
81
|
|
|
$stuffFor[1] = $edit->orig; |
82
|
|
|
$stuffFor[2] = $edit->orig; |
83
|
|
|
break; |
84
|
|
|
|
85
|
|
|
case 'change': |
86
|
|
|
$lookForTag = true; |
87
|
|
|
$stuffFor[1] = $edit->orig; |
88
|
|
|
$stuffFor[2] = $edit->final; |
89
|
|
|
break; |
90
|
|
|
|
91
|
|
|
case 'add': |
92
|
|
|
$lookForTag = true; |
93
|
|
|
$stuffFor[1] = null; |
94
|
|
|
$stuffFor[2] = $edit->final; |
95
|
|
|
break; |
96
|
|
|
|
97
|
|
|
case 'delete': |
98
|
|
|
$lookForTag = true; |
99
|
|
|
$stuffFor[1] = $edit->orig; |
100
|
|
|
$stuffFor[2] = null; |
101
|
|
|
break; |
102
|
|
|
} |
103
|
|
|
|
104
|
|
|
foreach ($stuffFor as $listName => $chunks) { |
105
|
|
|
if ($chunks) { |
106
|
|
|
foreach ($chunks as $item) { |
107
|
|
|
// $tagStack > 0 indicates that we should be tag-building |
108
|
|
|
if ($tagStack[$listName]) { |
109
|
|
|
$rechunked[$listName][sizeof($rechunked[$listName])-1] .= ' ' . $item; |
110
|
|
|
} else { |
111
|
|
|
$rechunked[$listName][] = $item; |
112
|
|
|
} |
113
|
|
|
|
114
|
|
|
if ($lookForTag |
115
|
|
|
&& !$tagStack[$listName] |
116
|
|
|
&& isset($item[0]) |
117
|
|
|
&& $item[0] == "<" |
118
|
|
|
&& substr($item, 0, 2) != "</" |
119
|
|
|
) { |
120
|
|
|
$tagStack[$listName] = 1; |
121
|
|
|
} elseif ($tagStack[$listName]) { |
122
|
|
|
if (substr($item, 0, 2) == "</") { |
123
|
|
|
$tagStack[$listName]--; |
124
|
|
|
} elseif (isset($item[0]) && $item[0] == "<") { |
125
|
|
|
$tagStack[$listName]++; |
126
|
|
|
} |
127
|
|
|
} |
128
|
|
|
} |
129
|
|
|
} |
130
|
|
|
} |
131
|
|
|
} |
132
|
|
|
|
133
|
|
|
// Diff the re-chunked data, turning it into maked up HTML |
134
|
|
|
$diff = new Diff($rechunked[1], $rechunked[2]); |
135
|
|
|
$content = ''; |
136
|
|
|
foreach ($diff->edits as $edit) { |
137
|
|
|
$orig = ($escape) ? Convert::raw2xml($edit->orig) : $edit->orig; |
138
|
|
|
$final = ($escape) ? Convert::raw2xml($edit->final) : $edit->final; |
139
|
|
|
|
140
|
|
|
switch ($edit->type) { |
141
|
|
|
case 'copy': |
142
|
|
|
$content .= " " . implode(" ", $orig) . " "; |
|
|
|
|
143
|
|
|
break; |
144
|
|
|
|
145
|
|
|
case 'change': |
146
|
|
|
$content .= " <ins>" . implode(" ", $final) . "</ins> "; |
147
|
|
|
$content .= " <del>" . implode(" ", $orig) . "</del> "; |
148
|
|
|
break; |
149
|
|
|
|
150
|
|
|
case 'add': |
151
|
|
|
$content .= " <ins>" . implode(" ", $final) . "</ins> "; |
152
|
|
|
break; |
153
|
|
|
|
154
|
|
|
case 'delete': |
155
|
|
|
$content .= " <del>" . implode(" ", $orig) . "</del> "; |
156
|
|
|
break; |
157
|
|
|
} |
158
|
|
|
} |
159
|
|
|
|
160
|
|
|
return self::cleanHTML($content); |
161
|
|
|
} |
162
|
|
|
|
163
|
|
|
/** |
164
|
|
|
* @param string|bool|array $content If passed as an array, values will be concatenated with a comma. |
165
|
|
|
* @return array |
166
|
|
|
*/ |
167
|
|
|
public static function getHTMLChunks($content) |
168
|
|
|
{ |
169
|
|
|
if ($content && !is_string($content) && !is_array($content) && !is_numeric($content) && !is_bool($content)) { |
|
|
|
|
170
|
|
|
throw new InvalidArgumentException('$content parameter needs to be a string or array'); |
171
|
|
|
} |
172
|
|
|
if (is_bool($content)) { |
173
|
|
|
// Convert boolean to strings |
174
|
|
|
$content = $content ? "true" : "false"; |
175
|
|
|
} |
176
|
|
|
if (is_array($content)) { |
177
|
|
|
$content = array_filter($content, 'is_scalar'); |
178
|
|
|
// Convert array to CSV |
179
|
|
|
$content = implode(',', $content); |
180
|
|
|
} |
181
|
|
|
|
182
|
|
|
$content = str_replace(array(" ", "<", ">"), array(" "," <", "> "), $content); |
183
|
|
|
$candidateChunks = preg_split("/[\t\r\n ]+/", $content); |
184
|
|
|
$chunks = []; |
185
|
|
|
for ($i = 0; $i < count($candidateChunks); $i++) { |
|
|
|
|
186
|
|
|
$item = $candidateChunks[$i]; |
187
|
|
|
if (isset($item[0]) && $item[0] == "<") { |
188
|
|
|
$newChunk = $item; |
189
|
|
|
while ($item[strlen($item)-1] != ">") { |
190
|
|
|
if (++$i >= count($candidateChunks)) { |
191
|
|
|
break; |
192
|
|
|
} |
193
|
|
|
$item = $candidateChunks[$i]; |
194
|
|
|
$newChunk .= ' ' . $item; |
195
|
|
|
} |
196
|
|
|
$chunks[] = $newChunk; |
197
|
|
|
} else { |
198
|
|
|
$chunks[] = $item; |
199
|
|
|
} |
200
|
|
|
} |
201
|
|
|
return $chunks; |
202
|
|
|
} |
203
|
|
|
} |
204
|
|
|
|