[3738] shorten_subject issue

[Özgür]

Code Select Expand

',shorten_subject($something['some'], 20),'

This code should work this

Code Select Expand

<a href="">arkadaslar özel kılıç...</a><br />
Here's how it works

Code Select Expand

<a href="">arkadaslar özel kılı�...</a><br />

� �   character non-utf8 and get xhtml error.
If i increasing 20 to 21 work fine,   looks "ç" but if twenty-first character is "ç" issue again..

Code Select Expand

',shorten_subject($something['some'], 21),'

I can't see any bug like this. I change my language files to original one but not work.

I can't validate my page..
System : 2.0 Rc1
Please look attacment.

[karlbenson]

moving to bug reports.

[Özgür]

I did not know and not sure this is bug. So I opened a topic 2.0 Help. Sorry and thanks. =)

[SleePy]

Are you using UTF-8?

[Özgür]

Are you using UTF-8?

Yes.

[Özgür]

I remember now..
I try to "Convert UTF8" in admin panel long time ago (I guess Rc1 released.) And Converter have a bug.. Maybe this is the problem.

Converter bug is fixed?

[SleePy]

It shouldn't be an issue with new posts, only old posts need the conversion.

We can try something.
Run This query In phpMyAdmin (What is phpMyAdmin?)

Code Select Expand


REPLACE INTO smf_settings VALUES ('disableEntityCheck', 1);

See if this works as it should now. shorten_subject appears to be using $smcFunc variables. So it should be UTF-8 safe. If this doesn't work, we will reverse that change.

[Özgür]

It shouldn't be an issue with new posts, only old posts need the conversion.

We can try something.
Run This query In phpMyAdmin (What is phpMyAdmin?)

Code Select Expand


REPLACE INTO smf_settings VALUES ('disableEntityCheck', 1);

See if this works as it should now. shorten_subject appears to be using $smcFunc variables. So it should be UTF-8 safe. If this doesn't work, we will reverse that change.

I run this code in myphpmadmin. And nothing change. And i reverse to "0".
Utf-8 character encoding database values how seen?

My messages seems like this (in phpmyadmin)

Code Select Expand

Ynt: Gitar Alıcam Yardım Lütfen

[Özgür]

Plus  i have this error in error_log file.
Please help me for this.

[SleePy]

Özgür´,

That first error tells me your SMF is missing some core checks.
I would suggest a reuploading all files from a large upgrade package (without the actual upgrade files). This will remove any modifications and custom themes.

Before you reapply your mods and themes. Test to see if new posts still have this issue. If you are still having the issues, then you can re apply mods. Otherwise re apply them 1 by 1, testing to see what mod breaks this.

[Özgür]

This topic moved but problem not fixed yet. I change my shorten_subject function.

Code Select Expand

// Shorten a subject + internationalization concerns.
function shorten_subject($subject, $len)
{
global $smcFunc;

// It was already short enough!
if ($smcFunc['strlen']($subject) <= $len)
return $subject;

// Shorten it by the length it was too long, and strip off junk from the end.
return mb_substr($subject, 0, $len, 'UTF-8') . '...';


}
But why smf function not correctly worked?

[karlbenson]

There I moved it back.

Have you tried what Sleepy suggested above?

Özgür´,

That first error tells me your SMF is missing some core checks.
I would suggest a reuploading all files from a large upgrade package (without the actual upgrade files). This will remove any modifications and custom themes.

Before you reapply your mods and themes. Test to see if new posts still have this issue. If you are still having the issues, then you can re apply mods. Otherwise re apply them 1 by 1, testing to see what mod breaks this.

[Özgür]

There I moved it back.

Have you tried what Sleepy suggested above?

Özgür´,

That first error tells me your SMF is missing some core checks.
I would suggest a reuploading all files from a large upgrade package (without the actual upgrade files). This will remove any modifications and custom themes.

Before you reapply your mods and themes. Test to see if new posts still have this issue. If you are still having the issues, then you can re apply mods. Otherwise re apply them 1 by 1, testing to see what mod breaks this.

Yes i hardly try actually.

Removed all files Sources files, Theme files etc. (not include settings.php) And i send clean smf 2.0 files. And send Upgrade files (upgrade sqls and upgrade.php) And i try upgrade again. Issue not solved. If you need anything (ssh, ftp, phpmyadmin or whatever) i will give.

[Özgür]

Solve is different function. Because problem is "substr" function for php side..

Code Select Expand

<?php
$ifade = 'Ne zaman seni düşünsem';
echo substr($ifade,0,20);
?>
Give us = > Ne zaman seni düş����
but mb_substr

Code Select Expand

<?php
$ifade = 'Ne zaman seni düşünsem';
echo mb_substr($ifade,0,20,'UTF-8');
?>
Give us to => Ne zaman seni düşüns

[k14]

Version(s) of SMF: SMF 2.0 RC1-1
Any non-English Language packs I have installed: polish-utf8
Am I using UTF-8? YES

Where the Error Occurred:
File: /Themes/default/Calendar.template.php
Line: 307
Is:
<td class="titlebg2" width="14%" align="center" ', $calendar_data['size'] == 'small' ? 'style="font-size: x-small;"' : '', '>', !empty($calendar_data['short_day_titles']) ? substr($txt['days'][$day], 0, 1) : $txt['days'][$day], '</td>';
Should be:
<td class="titlebg2" width="14%" align="center" ', $calendar_data['size'] == 'small' ? 'style="font-size: x-small;"' : '', '>', !empty($calendar_data['short_day_titles']) ? mb_substr($txt['days'][$day], 0, 1, $context['character_set']) : $txt['days'][$day], '</td>';

Additional Information:
In polish "Wednesday" is "Środa". Using short day names in calendar with substr() causes Środa to be truncated to one byte only, not to one character. This results in displaying the first byte of 2-byte utf8 sequence instead of "Ś".
I suggest changing all occurrences of substr() to mb_substr() in all files.

[Özgür]

Related shorten_subject issue

[karlbenson]

Finally got around to tracking this bug report inline with a dev team discussion topic.
http://dev.simplemachines.org/mantis/view.php?id=3738

[Özgür]

I glad to hear that.

[Elmacik]

A good catch. In deed using standart mb functions everywhere won't solve the issue PHP's standard multibyte functions doesn't seem to work with converted entities. For example if you use mb_substr() instead of substr, it will help a lot for UTF-8, but it will break entities. Because it counts every character by one even in entities. SMF will have to workaround that and I don't think it will be hard.

For example mb_substr('some text &copy; some other text', 0, 12'); returns "some text &c" normally. But it causes wrong displaying so it should return "some text &copy; " naturally and it will display correctly. (Just like the behaviour difference between normal sorting and natural sorting.) Its nothing to do with special entities, but some UTF-8 chars will turn into HTML entities too. For example some letter in Turkish alphabet as Daydreamer pointed out.

[†MavN†]

Still have issues in RC2 with utf-8.
For example, calendar's short day titles

Code Select Expand

<th class="titlebg2 days" scope="col" ', $calendar_data['size'] == 'small' ? 'style="font-size: x-small;"' : '', '>', !empty($calendar_data['short_day_titles']) ? substr($txt['days'][$day], 0, 1) : $txt['days'][$day], '</th>';

Should be replaced with

Code Select Expand

<th class="titlebg2 days" scope="col" ', $calendar_data['size'] == 'small' ? 'style="font-size: x-small;"' : '', '>', !empty($calendar_data['short_day_titles']) ? $smcFunc['substr']($txt['days'][$day], 0, 1) : $txt['days'][$day], '</th>';