National Characters in names in Search.

Submitted by Axolotl on Mon, 12/24/2012 - 00:38
Forums: 
Soulseek General Discussion

Hi, I'm not sure if it's a bug report, feature request or just a questuion.
I'm not sure exactly but AFAIR previously soulseek behaviour was differnet.
For example when you type in search Bjorn Stabi it will be good to show also Björn Ståbi in results. And as far as i remeber previously soulseek really did such thing..or maybe not? If not...it's very good to be such thing in SS,

Also when you try to search with such characters it also don't give results, f.e. "björn ståbi & ole hjorth" don't gives results, but when you try to search ole hjorth it finds it.

So, it's big problem to find such names now.....

Sorry for my english....

Submitted by Nir on Sat, 12/29/2012 - 14:00.

Hmm, I think you're right. There might have been something in the original client that was indexing shared files using tokens with converted accents. I'd have to look into how easy it would be to do with same with SoulseekQt using Unicode.

Submitted by sysmaster01 on Fri, 01/04/2013 - 08:00.

I suddenly remember there was a related inquiry (though worded as a feature request) in the other forum not long ago ...
http://www.soulseekqt.net/news/node/1433

Note, I didn't say "similar" but "related" for a reason. Yes the other one is about replacing characters deemed "illegal" in certain OS. However, both inquiries do have one thing in common: they will have to make use of the same technique: the "on-the-fly replacing" of characters in file/folder names.

Anyways, the way how results are displayed in SlskQt is 100% bug-free AFAICS, since the "problem" comes from people still using the old non-Unicode (!) SlskNG.
You would get the same results if you had a ISO-8859-1/-15 database and tried to view accented or umlaut characters in a full-Unicode-configured MySQL server setup.

Nir, whatever you do to circumvent this "problem" will surely be artificial, i. e. "falsifying" results to comply with the shortcomings of former SlskNG builds; commonly called a "workaround". :)

Submitted by Axolotl on Sun, 01/06/2013 - 07:07.

Hmm...maybe i understand wrong, but i was speaking not only about system code page differences problems.
I was speaking also about search results in which some characters will be "equalling".
For example æ = ae; ð = d; à,á,â,ã,ä,å = a (and a also = à,á,â,ã,ä,å) and so on...all these "a" variatns better to be equal to search engine (maybe in optional way).
This feature is very important not only with difference in code pages, but simply because many people have their artist folder names and artist tags with national characters and some with such simplified latinic symbols, f.e. some people name folder "Árstíðir" and some simplify it to Arstidir...so it will be good if search will find and return both variants to both variants of search requests (simplified and national).

By the way...such feature do exist in some software. For example eMule have it, and even Foobar2000 (Playlist search) have such feature.