harukin/core
Archived
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix html2markdown() and re-enable previously failing tests

Browse Source
This commit is contained in:
Mario Vavti 2018-10-30 13:55:08 +01:00
parent 2192ea6fdf
commit 0b371c8103
2 changed files with 6 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
include/markdown.php
View File

@ -248,20 +248,12 @@ function bb_to_markdown($Text, $options = []) {
// Convert it to HTML - don't try oembed // Convert it to HTML - don't try oembed
$Text = bbcode($Text, [ 'tryoembed' => false ]); $Text = bbcode($Text, [ 'tryoembed' => false ]);
// Markdownify does not preserve previously escaped html entities such as <> and &.
//$Text = str_replace(array('&lt;','&gt;','&amp;'),array('&_lt_;','&_gt_;','&_amp_;'),$Text);
// Now convert HTML to Markdown // Now convert HTML to Markdown
$Text = html2markdown($Text); $Text = html2markdown($Text);
//html2markdown adds backslashes infront of hashes after a new line. remove them //html2markdown adds backslashes infront of hashes after a new line. remove them
$Text = str_replace("\n\#", "\n#", $Text); $Text = str_replace("\n\#", "\n#", $Text);
// It also adds backslashes to our attempt at getting around the html entity preservation for some weird reason.
//$Text = str_replace(array('&\\_lt\\_;','&\\_gt\\_;','&\\_amp\\_;'),array('&lt;','&gt;','&amp;'),$Text);
// If the text going into bbcode() has a plain URL in it, i.e. // If the text going into bbcode() has a plain URL in it, i.e.
// with no [url] tags around it, it will come out of parseString() // with no [url] tags around it, it will come out of parseString()
// looking like: <http://url.com>, which gets removed by strip_tags(). // looking like: <http://url.com>, which gets removed by strip_tags().
@ -298,7 +290,8 @@ function html2markdown($html,$options = []) {
$internal_errors = libxml_use_internal_errors(true); $internal_errors = libxml_use_internal_errors(true);
$environment = Environment::createDefaultEnvironment($options); $environment = new Environment($options);
$environment->createDefaultEnvironment();
$environment->addConverter(new TableConverter()); $environment->addConverter(new TableConverter());
$converter = new HtmlConverter($environment); $converter = new HtmlConverter($environment);

8
tests/unit/includes/MarkdownTest.php
View File

@ -52,12 +52,12 @@ class MarkdownTest extends UnitTestCase {
' &nbsp;', ' &nbsp;',
'' ''
], ],
/* it is not clear why this test fails
'strong, b, em, i, bib' => [ 'strong, b, em, i, bib' => [
'<strong>strong</strong> <b>bold</b> <em>em</em> <i>italic</i> <b>bo<i>italic</i>ld</b>', '<strong>strong</strong> <b>bold</b> <em>em</em> <i>italic</i> <b>bo<i>italic</i>ld</b>',
'**strong** **bold** _em_ _italic_ **bo_italic_ld**' '**strong** **bold** _em_ _italic_ **bo_italic_ld**'
], ],
*/
'empty tags' => [ 'empty tags' => [
'text1 <b></b> text2 <i></i>', 'text1 <b></b> text2 <i></i>',
'text1 text2' 'text1 text2'
@ -118,12 +118,12 @@ class MarkdownTest extends UnitTestCase {
'<code>&lt;p&gt;HTML text&lt;/p&gt;</code>', '<code>&lt;p&gt;HTML text&lt;/p&gt;</code>',
'`<p>HTML text</p>`' '`<p>HTML text</p>`'
], ],
/* it is not clear why this test fails
'pre' => [ 'pre' => [
'<pre> line with spaces </pre>', '<pre> line with spaces </pre>',
'` line with spaces `' '` line with spaces `'
], ],
*/
'div p' => [ 'div p' => [
'<div>div</div><div><p>p</p></div>', '<div>div</div><div><p>p</p></div>',
"<div>div</div><div>p\n\n</div>" "<div>div</div><div>p\n\n</div>"