moodle/moodle
3
0
Fork 0
mirror of https://github.com/moodle/moodle.git synced 2025-08-04 08:26:37 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

MDL-17542 lib/html2text: update to latest upstream version (r2374)

Browse source 
from http://trac.roundcube.net/log/trunk/roundcubemail/program/lib/html2text.php
This commit is contained in:
fmarier 2009-05-22 02:10:22 +00:00
parent db26acd41d
commit ea2678d8b9
3 changed files with 51 additions and 50 deletions
Download patch file Download diff file Expand all files Collapse all files

93
lib/html2text.php
View file

@ -176,7 +176,6 @@ class html2text
'/&(bull|#149|#8226);/i', // Bullet '/&(bull|#149|#8226);/i', // Bullet
'/&(pound|#163);/i', // Pound sign '/&(pound|#163);/i', // Pound sign
'/&(euro|#8364);/i', // Euro sign '/&(euro|#8364);/i', // Euro sign
'/&[^&;]+;/i', // Unknown/unhandled entities
'/[ ]{2,}/' // Runs of spaces, post-handling '/[ ]{2,}/' // Runs of spaces, post-handling
); );
@ -202,8 +201,8 @@ class html2text
"\n\n", // <ol> and </ol> "\n\n", // <ol> and </ol>
"\t* \\1\n", // <li> and </li> "\t* \\1\n", // <li> and </li>
"\n\t* ", // <li> "\n\t* ", // <li>
"\n-------------------------\n", // <hr> "\n-------------------------\n", // <hr>
"\n\n", // <table> and </table> "\n\n", // <table> and </table>
"\n", // <tr> and </tr> "\n", // <tr> and </tr>
"\t\t\\1\n", // <td> and </td> "\t\t\\1\n", // <td> and </td>
' ', // Non-breaking space ' ', // Non-breaking space
@ -220,7 +219,6 @@ class html2text
'*', '*',
'£', '£',
'EUR', // Euro sign. € ? 'EUR', // Euro sign. € ?
'', // Unknown/unhandled entities
' ' // Runs of spaces, post-handling ' ' // Runs of spaces, post-handling
); );
@ -249,11 +247,11 @@ class html2text
* @see $pre_replace * @see $pre_replace
*/ */
var $pre_search = array( var $pre_search = array(
"/\n/", "/\n/",
"/\t/", "/\t/",
'/ /', '/ /',
'/<pre[^>]*>/', '/<pre[^>]*>/',
'/<\/pre>/' '/<\/pre>/'
); );
/** /**
@ -264,11 +262,11 @@ class html2text
* @see $pre_search * @see $pre_search
*/ */
var $pre_replace = array( var $pre_replace = array(
'<br>', '<br>',
'&nbsp;&nbsp;&nbsp;&nbsp;', '&nbsp;&nbsp;&nbsp;&nbsp;',
'&nbsp;', '&nbsp;',
'', '',
'' ''
); );
/** /**
@ -344,10 +342,10 @@ class html2text
if ( !empty($source) ) { if ( !empty($source) ) {
$this->set_html($source, $from_file); $this->set_html($source, $from_file);
} }
$this->set_base_url(); $this->set_base_url();
$this->_do_links = $do_links; $this->_do_links = $do_links;
$this->width = $width; $this->width = $width;
} }
/** /**
@ -361,10 +359,10 @@ class html2text
function set_html( $source, $from_file = false ) function set_html( $source, $from_file = false )
{ {
if ( $from_file && file_exists($source) ) { if ( $from_file && file_exists($source) ) {
$this->html = file_get_contents($source); $this->html = file_get_contents($source);
} }
else else
$this->html = $source; $this->html = $source;
$this->_converted = false; $this->_converted = false;
} }
@ -465,16 +463,19 @@ class html2text
$text = trim(stripslashes($this->html)); $text = trim(stripslashes($this->html));
// Convert <PRE> // Convert <PRE>
$this->_convert_pre($text); $this->_convert_pre($text);
// Replace known html entities
$text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
// Run our defined search-and-replace // Run our defined search-and-replace
$text = preg_replace($this->search, $this->replace, $text); $text = preg_replace($this->search, $this->replace, $text);
$text = preg_replace_callback($this->callback_search, array('html2text', '_preg_callback'), $text); $text = preg_replace_callback($this->callback_search, array('html2text', '_preg_callback'), $text);
// Replace known html entities
$text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
// Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
$text = preg_replace('/&[^&;]+;/i', '', $text);
// Strip any other HTML tags // Strip any other HTML tags
$text = strip_tags($text, $this->allowed_tags); $text = strip_tags($text, $this->allowed_tags);
@ -521,9 +522,9 @@ class html2text
$this->_link_count++; $this->_link_count++;
$this->_link_list .= "[" . $this->_link_count . "] $link\n"; $this->_link_list .= "[" . $this->_link_count . "] $link\n";
$additional = ' [' . $this->_link_count . ']'; $additional = ' [' . $this->_link_count . ']';
} elseif ( substr($link, 0, 11) == 'javascript:' ) { } elseif ( substr($link, 0, 11) == 'javascript:' ) {
// Don't count the link; ignore it // Don't count the link; ignore it
$additional = ''; $additional = '';
// what about href="#anchor" ? // what about href="#anchor" ?
} else { } else {
$this->_link_count++; $this->_link_count++;
@ -546,11 +547,10 @@ class html2text
*/ */
function _convert_pre(&$text) function _convert_pre(&$text)
{ {
while(preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) while(preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
{ $result = preg_replace($this->pre_search, $this->pre_replace, $matches[1]);
$result = preg_replace($this->pre_search, $this->pre_replace, $matches[1]); $text = preg_replace('/<pre[^>]*>.*<\/pre>/ismU', '<div><br>' . $result . '<br></div>', $text, 1);
$text = preg_replace('/<pre[^>]*>.*<\/pre>/ismU', '<div><br>' . $result . '<br></div>', $text, 1); }
}
} }
/** /**
@ -562,18 +562,17 @@ class html2text
*/ */
function _preg_callback($matches) function _preg_callback($matches)
{ {
switch($matches[1]) switch($matches[1]) {
{ case 'b':
case 'b': case 'strong':
case 'strong': return $this->_strtoupper($matches[2]);
return $this->_strtoupper($matches[2]); case 'hr':
case 'hr': return $this->_strtoupper("\t\t". $matches[2] ."\n");
return $this->_strtoupper("\t\t". $matches[2] ."\n"); case 'h':
case 'h': return $this->_strtoupper("\n\n". $matches[2] ."\n\n");
return $this->_strtoupper("\n\n". $matches[2] ."\n\n"); case 'a':
case 'a': return $this->_build_link_list($matches[3], $matches[4]);
return $this->_build_link_list($matches[3], $matches[4]); }
}
} }
/** /**
@ -585,10 +584,10 @@ class html2text
*/ */
function _strtoupper($str) function _strtoupper($str)
{ {
if (function_exists('mb_strtoupper')) if (function_exists('mb_strtoupper'))
return mb_strtoupper($str); return mb_strtoupper($str);
else else
return strtoupper($str); return strtoupper($str);
} }
} }

5
lib/html2text_readme.txt Normal file
View file

@ -0,0 +1,5 @@
html2text.php is an unmodified copy of a file shipped with the RoundCube project:
http://trac.roundcube.net/log/trunk/roundcubemail/program/lib/html2text.php
-- Francois Marier <francois@catalyst.net.nz> 2009-05-22

3
lib/weblib.php
View file

@ -1986,9 +1986,6 @@ function html_to_text($html) {
$h2t = new html2text($html); $h2t = new html2text($html);
$result = $h2t->get_text(); $result = $h2t->get_text();
// html2text does not fix HTML entities so handle those here.
$result = trim(html_entity_decode($result, ENT_NOQUOTES, 'UTF-8'));
return $result; return $result;
} }