Textpattern CMS support forum

You are not logged in. Register | Login | Help

#11 2005-11-01 22:10:15

Jeremie
Member
From: Provence, France
Registered: 2004-08-11
Posts: 1,578
Website

Re: [issue] 4.0.1: search and case of non-ascii letters

Another related – I think – issue, and maybe a bigger one is the case of the apostrophe (‘). It’s automaticaly transformed by Textile to the right character, it’s heavily used in several languages (including french and italian) but how the search is handling it since no keyword or even special driver – to my knowledge – are able to output it.

Put it another way: Textile transform the single quote character <code>’</code> into a real unicode apostrophe ‘ , but no one using the search tool will use the real one they will all use the single quote character.

I’ve done some test.. well the results are strange… TXP does find some articles with words with apostrophe, but find another set of articles if the real apostrophe character is copied/pasted. Really strange.

Offline

#12 2005-11-01 23:42:03

zem
Developer emeritus
From: Melbourne, Australia
Registered: 2004-04-08
Posts: 2,579
Website

Re: [issue] 4.0.1: search and case of non-ascii letters

Textpattern searches the text before Textile is applied, exactly as it’s entered in the body textarea.

The difference might be caused by two different encodings of the same character (ASCII 0×27 vs. Unicode U+0027), or by two different characters that might appear similar on screen (U+0027 vs U+2019).

At any rate, MySQL’s fulltext indexing isn’t really designed to include puncutation in searches.


Alex
tstate

Offline

#13 2005-11-02 15:53:18

Jeremie
Member
From: Provence, France
Registered: 2004-08-11
Posts: 1,578
Website

Re: [issue] 4.0.1: search and case of non-ascii letters

Noted about the pre-Textile search, and the similar characters. I will investigate more on that topic, thanks.

Offline

Board footer

Powered by FluxBB