Wildcards not working properly in non-english letters

[Tapio]

I found inconsistencies when using wildcards in words that have non-english letters (ä,ö). Examples (for clarity the search terms are in capitals):

• ÄITI -> 71 hits, but ?ITI -> no hits and ÄITI* -> only 57 hits (why less?)
• ÄITI-HENKI -> 87 hits, but ÄITI-HE* -> no hits
• ÖLJY -> 2 hits, but ?LJY -> no hits
• SILMÄ -> 15 hits, but SILM? -> no hits
• KÄSIN -> 46 hits, but K?SIN -> no hits
• etc.

Still any words having english alphabets only are found to work correctly in examples like above.
All html files and the Zoom configuration are defined for UTF-8.
I use PHP platform and Zoom 8, build 1026.

[David]

Looks like we might be using the "preg_match" PHP function to do this wild card matching.
It might not support UTF-8 by default.

We'll dig a bit deeper and then post an update.

[Simon (PassMark)]

Made some changes to the php search script - can you replace existing search.php with the attached

[Tapio]

Wau! This fix made a change. A huge improvement in wildcard searches, when there are non-english letters involved. Seem to work perfect.

One more thing I noticed, was in exclusion word ("-word"). That term does not honour any wildcard. But this has nothing to do with non-english letters, it applies always. So maybe this is a feature, not a defect in design?

Many thanks!

[David]

Great that it fixed it.

Yes, I don't think negative search terms with wild cards have ever been supported.

We did a study some years ago. 95%+ of people never do anything more complex that one or two word searches. Very very few people ever use the advanced search syntax.