Author Topic: Unicode?? (Read 15800 times)

geoff915 · « **on:** November 14, 2008, 02:31 PM »

I can't seem to find the issue of Unicode having been addressed previously. Is there any chance that we will see Unicode support sometime in the future? I consider FARR to be vastly superior and preferable to the alternatives I've seen... except that one issue.

Should I hold out hope, or resign myself to manually calling up Launchy on the occasions when I have to use international characters?

mouser · « **Reply #1 on:** November 14, 2008, 03:23 PM »

I'm afraid the bad news is that unicode support for FARR probably won't come anytime soon. Launchy may be your best bet for such things for the time being.

I'm definitely not ruling out unicode in the future -- there has been some discussion about rewriting FARR for a future version, and if i do that, i will make it unicode compatible -- i just don't want to try to patch it into the current codebase.

duchamp · « **Reply #2 on:** November 16, 2008, 09:36 AM »

Thank you, Mouser.
We are waiting for the unicode support.

echo · « **Reply #3 on:** November 22, 2008, 08:48 AM »

Did I miss something ?

When the term "UNICODE" is used - do you mean entering foreign language characters into the FARR search field ?
I ask because - I do it in the present - and it works
So I must be missing something here.

I will appreciate some clarification for my own education

f0dder · « **Reply #4 on:** November 22, 2008, 09:12 AM »

echo: yes, UNICODE means - basically - "foreign characters". SOME already work in FARR, but it's limited to the codepage you're running. Ie., with a Danish codepage I can enter "æøåÆØÅäëöüï" and stuff like that without problems, but FARR fails miserably for Chinese, Russian, etc.

tranglos · « **Reply #5 on:** November 22, 2008, 11:58 AM »

echo: yes, UNICODE means - basically - "foreign characters". SOME already work in FARR, but it's limited to the codepage you're running. Ie., with a Danish codepage I can enter "æøåÆØÅäëöüï" and stuff like that without problems, but FARR fails miserably for Chinese, Russian, etc.
-f0dder (November 22, 2008, 09:12 AM)

This is something I don't quite get. Assuming FARR is used in its basic function of looking up files, do you need Unicode to enter the name of any file in your filesystem? I can't quite imagine having files named using Chinese characters on a Central European version of Windows, for example. How is this useful (or even possible) in practice?

(It is of course another thing when using FARR with plugins, search Google or Firefox bookmarks etc. I'm not negating the usefulness of Unicode in general, just wondering about the filesystems.)

skajfes · « **Reply #6 on:** November 22, 2008, 12:10 PM »

This is something I don't quite get. Assuming FARR is used in its basic function of looking up files, do you need Unicode to enter the name of any file in your filesystem? I can't quite imagine having files named using Chinese characters on a Central European version of Windows, for example. How is this useful (or even possible) in practice?
-tranglos (November 22, 2008, 11:58 AM)

Yes, but what if you have Chinese version of Windows? Or in my case(i am from Croatia) when you have filenames with special characters like "šđčćž" which are quite often in filenames around here.

Crush · « **Reply #7 on:** November 22, 2008, 12:32 PM »

I think unicode (16-Bit WChar) makes some problems. It uses more space and therefore you have to search slower through strings than Ascii. UTF8-24 is even more slower to decode/search. If you decide to use unicode it´s simplier to jump over directly to UTF32 that has no problems with all existing characters.

f0dder · « **Reply #8 on:** November 22, 2008, 12:51 PM »

tranglos: tbh, I don't have a lot of chinese or russian files on my computer

- but iirc some non-English users had trouble with FARR, so perhaps there are issues when only searching the local codepage, with those "richer" languages.

Crush: 1990 was 18 years ago - utf parsing is only a problem if you do really speed-intensive stuff. And FARR is often disk-bound anyway. Might as well stick with utf-16 (which is what windows NT API internally uses) and forget about UCS-4/UTF-32, and conserve a bit of memory.

tranglos · « **Reply #9 on:** November 22, 2008, 03:13 PM »

Yes, but what if you have Chinese version of Windows? Or in my case(i am from Croatia) when you have filenames with special characters like "šđčćž" which are quite often in filenames around here.
-skajfes (November 22, 2008, 12:10 PM)

That makes sense, thanks

Although I'm in a very similar position (Poland) and I've seen no problems using FARR with Polish diacritic characters, with one exception, but it's unrelated to the codepage.

I wonder though how adding Unicode support to FARR would help in situations such as "X language filename on Y language system", since as far as I know filenames are not stored in Unicode on Windows - or are they? I've seen "unicode filenames" mentioned here and there, but Windows APIs return ANSI filenames, at least on XP.

Now the the exception (hope mouser sees it FARR uses Ctrl+Alt+O to trigger the Options dialog box. This blocks me from entering the "ó" character, a Polish diacritic. More precisely, the character does appear correctly in the search box, but the Options dialog box pops up as well, preventing me from typing any further until I dismiss it. It's a frequent gotcha when assigning Ctrl+Alt combinations, since the right Alt key (AltGR) on many country-specific keyboard maps to Ctrl+Alt. So for example any program that swallows Ctrl+Alt plus a, c, e, l, n, o, x or z prevents me from typing in Polish.

f0dder · « **Reply #10 on:** November 22, 2008, 03:18 PM »

NT has been fully unicode since the start - only 9x windows versions have limited unicode support. NTFS is also native unicode, and even FAT supports it (though it's an evil hack... using multiple filename descriptor entries, chaining them, etc). Windows APIs dealing with strings have two versions, btw, *A and *W - ANSI and WIDE/Unicode. ANSI versions internally convert to unicode, call the *W variant, and convert back to ANSI.

Lashiec · « **Reply #11 on:** November 26, 2008, 09:18 AM »

This is something I don't quite get. Assuming FARR is used in its basic function of looking up files, do you need Unicode to enter the name of any file in your filesystem? I can't quite imagine having files named using Chinese characters on a Central European version of Windows, for example. How is this useful (or even possible) in practice?
-tranglos (November 22, 2008, 11:58 AM)

Even for those not using Unicode characters, it makes sense. For example, I have several songs named with their Japanese name (those clever guys at MusicBrainz...), and when searching for the name of the artists (which use ASCII characters), I'll get these funny squares instead of the actual titles.

czb · « **Reply #12 on:** November 26, 2008, 10:41 AM »

Yes.. unicode is a must

I wonder if "fake" unicode would be possible. By that I mean that all input locations of FARR would convert unicode chars to something like utf01235 and store it internaly that way and on all output locations it would convert back.

Simple and maybe it would work, what do you think?

f0dder · « **Reply #13 on:** November 26, 2008, 11:13 AM »

czb: that'd be probably about the same amount of work as doing a full unicode version. Thing is, getting file search etc. in unicode isn't that bad, but iirc FARR currently uses some 3rd-party controls that don't have unicode support.

mouser · « **Reply #14 on:** November 26, 2008, 03:56 PM »

F0dder is basically summing up the debate.. which is the kind of debate that all such projects have to go through periodically.. which is:

At what point should you do a rewrite, vs. trying to patch in major changes.

Right now FARR is showing it's age -- there are soo many things that were learned over the SEVEN years since the first release of FARR, that it would really be nice to rewrite it from scratch, but that's a big project.

But adding something like unicode is likely to cause all sort of rippling effects in terms of plugins, etc., that it would be hard to patch something like that in.

HOWEVER, as czb suggested, it may be possible to come up with some kind of minor changes that would solve 90% of the problem while being short of converting to true unicode support.. and that might be worth looking into.

f0dder · « **Reply #15 on:** November 26, 2008, 04:11 PM »

mouser: if you could find a listview (or whatever it was) component that supports unicode (as well as ANSI), then you could probably keep FARR in ANSI, but do file traversal in unicode... that hopefully wouldn't break plugins.

Armando · « **Reply #16 on:** September 17, 2009, 05:20 PM »

mouser: if you could find a listview (or whatever it was) component that supports unicode (as well as ANSI), then you could probably keep FARR in ANSI, but do file traversal in unicode... that hopefully wouldn't break plugins.
-f0dder (November 26, 2008, 04:11 PM)

[sorry for resurrecting this old thread]

Any such thing possible (mouser) ? That would probably still be a lot of work. But also worth it IF many users use "foreign characters" in their FS.

skajfes · « **Reply #17 on:** September 18, 2009, 06:55 PM »

I don't know about you guys, but this solved all of my language related issues with FARR (and some other apps too). Go to Regional and Language options in Control Panel - select the Advanced tab, and under the Language for non-unicode programs select your language. Press OK.

mouser · « **Reply #18 on:** September 18, 2009, 06:58 PM »

now that is an interesting discovery!!

Armando · « **Reply #19 on:** September 18, 2009, 09:29 PM »

Interesting, yes. Windows is full of little things like that.
However this didn't change anything to my character problems...

Screenshot - 2009-09-18 , 22_23_18.png

skajfes · « **Reply #20 on:** September 19, 2009, 04:54 AM »

Have you tried setting font overrides in FARR? It is in Settings - Font overrides. Be sure to select script that suites you. I chose central european. Although I've tried playing with theese settings now and got no difference, but try it anyway.

Armando · « **Reply #21 on:** September 19, 2009, 12:10 PM »

Thanks skajfes!
Unfortunately it still doesn't change anything...
I cannot really choose any other option than "Western" anyways... Since I speak French.