Truncate is cutting within a tag

[laoneo]

Truncating a string is not working 100% when attributes are set. The following code is producing incorrect HTML output Some text is <div..., even the doc comment says it closes open tags.

JHtml::_('string.truncate', 'Some text is <div class="test"><span><i>Plain</i></span></div>', 16);

Probably the whole function needs to be rewritten, using a DOM parser.

Steps to reproduce the issue

In the file tests/unit/suites/libraries/cms/html/JHtmlStringTest.php after line 210, add the following code

'Dont show open tags' => array(
                'Some text is <div class="test"><span><i>Plain</i></span></div>',
                20,
                true,
                true,
                'Some text is ',
            ),

Run phpunit tests/unit/suites/libraries/cms/html/JHtmlStringTest.php.

Expected result

Unit test will run trough.

Actual result

1) JHtmlStringTest::testTruncate with data set "HTML not allowed, no split" ('Some text is <div class="test...</div>', 20, true, true, 'Some text is ')
Failed asserting that two strings are equal.
--- Expected
+++ Actual
@@ @@
-'Some text is '
+'Some text is <div...'

System information (as much as possible)

Joomla 3.6.2 on PHP 7.

[rdeutz]

I have made a fix for this years ago but nobody gave a fu?? and chose to discuss it to dead. Will check if I can find the fix. The problem is the space within the tag, if that is the place to spilt then anything goes south.

[laoneo]

Hope you will find the fix again 😬

[brianteeman]

Do you mean #978

On 6 September 2016 at 13:47, Robert Deutz notifications@github.com wrote:

I have made a fix for this years ago but nobody gave a fu?? and chose to
discuss it to dead. Will check if I can find the fix. The problem is the
space within the tag, if that is the place to spilt then anything goes
south.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#11934 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABPH8RQuIBobUR9k5OWHrvHUDM43MyrGks5qnWDlgaJpZM4J05V3
.

Brian Teeman
Co-founder Joomla! and OpenSourceMatters Inc.
http://brian.teeman.net/

[brianteeman]

or this one #1665

On 6 September 2016 at 13:52, Brian Teeman brian@teeman.net wrote:

Do you mean #978

On 6 September 2016 at 13:47, Robert Deutz notifications@github.com
wrote:

I have made a fix for this years ago but nobody gave a fu?? and chose to
discuss it to dead. Will check if I can find the fix. The problem is the
space within the tag, if that is the place to spilt then anything goes
south.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#11934 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABPH8RQuIBobUR9k5OWHrvHUDM43MyrGks5qnWDlgaJpZM4J05V3
.

Brian Teeman
Co-founder Joomla! and OpenSourceMatters Inc.
http://brian.teeman.net/

Brian Teeman
Co-founder Joomla! and OpenSourceMatters Inc.
http://brian.teeman.net/

[rdeutz]

The "endless" story is here on joomlacode 30552 :-) and yes #1665 seems to be the last one. It only happens from time to time more then 3 years someone faced the same problem. Going to make a patch soon

[laoneo]

@rdeutz thank you!
@brianteeman did you remember the ids without search 😄 ?

[rdeutz]

Tried it, but old patch doesn't work, I need a bit more time to check and fix

[brianteeman]

@brianteeman did you remember the ids without search 😄 ?

of course

[eshiol]

JHtmlString::truncateComplex solved the issue #12612 because it is more reliable than JHtmlString::truncate. What do you think to change JHtmlString::truncate so you automatically use JHtmlString::truncateComplex when allowHtml is true?

[laoneo]

JHtmlString::truncateComplex is also not solving the issue as the error happens there as well.

[ghost]

is this Issue similar to #11835, #13594?

[laoneo]

@franz-wohlkoenig no. This one here is about real truncating HTML code by removing letters and keeping the HTML structure valid like closing opening tags after truncate.

[ghost]

thanks for info, @laoneo

[rdeutz]

closing because we have a PR #14907