Issue: Automatic linker is case-sensitive with non-latin characters

Issue Tools
- View Changes

November 24, 2010 11:45 AM
Mikele

New Member
Automatic linker is case-sensitive with non-latin characters

Automatic linker is case-sensitive with non-latin characters

Found that when using the Cyrillic characters automatic linker only handles the words in lowercase. At the same time, if the article name is in English, then word processing is case-insensitive.

Examples:

Code:

Article names: Docs | Масло Words that may be linked: docs Docs DoCs | масло Words that do not become links: - | Масло мАсло

Issue Details

Issue Number 2105

Issue Type Bug

Project VaultWiki 3.x Series

Category Non-Wiki Threads / Posts

Status Fixed

Priority 3 - Loss of Functionality

Affected Version 3.0.6

Fixed Version 3.0.8

Milestone VaultWiki 3.0.8

Software DependencyAny

Users able to reproduce bug 0

Users unable to reproduce bug 0

Attachments 0

Assigned Users (none)

Tags (none)

November 24, 2010 12:26 PM

pegasus

VaultWiki Team

PHP has to be compiled with Multi-Byte String support, the forum has to be using UTF-8 encoding, and UTF-8 support must be enabled in AdminCP > Settings > Options > VaultWiki: Server Settings > Enable UTF-8 Support for URLs? for case-insensitivity to work properly. If any one of these is not true, VaultWiki will not be able to determine that any non-ASCII characters are the same in different cases.

Reply
November 24, 2010 12:33 PM

Mikele

New Member

Yes, PHP is compiled with the mbstring and also the iconv modules (http://forum.rotarycars.ru/tmp231.php - my phpinfo), and I set Yes on "Enable UTF-8 Support for URLs?".

Reply
November 24, 2010 12:45 PM

Mikele

New Member

Also tried to change the following mbstring settings:
mbstring.language = Russian
mbstring.internal_encoding = UTF-8

+ cleared the cache and rebuilt URLs. Did not help

Reply
November 24, 2010 12:52 PM

pegasus

VaultWiki Team

Okay I will do some tests and see if I can reproduce the issue using the examples you posted in the first post.

Edit: Is the a in the ~macro~ article title an ASCII a? If it is, then it is expecting ASCII A as its upper-case version, and there's nothing we can really do about this. If you want both to be autolinked, you would have to create a redirect-page that has the Cyrillic variant.

Reply
November 24, 2010 1:21 PM

Mikele

New Member

As far as I understand your question... we always use UTF-8 encoding. To be sure, I did a dump of hexadecimal character values of the original string from the database, as well as the same characters in ASCII.
UTF-8 String HEX-codes: 041C 0430 0441 043B 043E (got this directly from database)
ASCII String HEX-codes: CC E0 F1 EB EE (this how it will be in a windows-1251 cyrillic encoding)

Reply
November 24, 2010 1:28 PM

pegasus

VaultWiki Team

Okay, thanks. This answers my question. I will do some testing with the Cyrillic lower-case a and try to figure out why it's not working for you.

мАсло

Reply
December 7, 2010 1:41 AM

pegasus

VaultWiki Team

Just bumping this so it gets addressed in 3.0.8.

Reply
December 7, 2010 2:18 PM

pegasus

VaultWiki Team

Fixed by turning on UTF-8 mode (adding the 'u' modifier to the autolink regex). According to the PHP doc, UTF-8 mode is available in all the PHP versions VaultWiki requires, so I doubt additional checking would be necessary. Will be included in the next release.

Reply
December 7, 2010 6:23 PM

Mikele

New Member

Great news! Thank you, Pegasus!

Reply

+ Reply

All times are GMT -4. The time now is 2:58 PM.

This site uses cookies to help personalize content, to tailor your experience, and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.

Learn more… Accept Remind me later

Welcome to VaultWiki.org, home of the wiki add-on for vBulletin and XenForo!

Issue: Automatic linker is case-sensitive with non-latin characters

Issue Tools