UTF-8 in Slackware -what's missing?
 

Hello,

With 2.6.24 in the latest -current, UTF-8 seems to be one step closer to Slackware.

Does anyone know what are the items outstanding or problems to watch out for in seeking to make Slackware 100% UTF-8 ?

I know the man pages used to be a problem, but have found no recent information on that subject. Does anyone else have input on the progress on UTF-8 man-pages

Are there other items one should be aware of before starting the conversion?

(As Slackware keeps gaining momentum among non-American Linux-users, this gets more relevant every day)

-y1

Well, there was/is a problem with hplip GUI tools because they expect UTF8 input (see this thread ). rworkman is aware of it and the work-around, though. That is the only i18n issue I am aware of.

EDIT: The hp-tools workround is now mentioned in the slackware-current CHANGES_AND_HINTS.txt

Midnight Commander needs a patch to display utf8 characters, and it also needs a utf8-patched slang2.
Further I had some problems with kmail authenticating to a iso-8859-15 server, with utf8 enabled locally.
But I don't know, if that problem was caused by kmail or the server.

I'm using UTF-8 on Slackware for over 2 years now - no problem at all.

The problem with manpages is the support for UTF-8 in groff (AFAIK) and the last time I checked, it couldn't display Thai script, for example.

Most application dealing with IDN (which is not UTF-8, but if we're talking i18n here... :) I use work also perfectly fine. (mutt, Thunderbird, Firefox, qmail to name a few working apps...)

Input methods with scim do fine, printing works, Perl's got some nice features and support for Unicode stuff..

I'm missing nothing, actually.

(And as a non-american Slackware user: I'm using Slackware since '94. :)

I've switched to UTF8 now myself and take back my comment on Midnight Commander. I don't know what all this complaints about missing UTF8 support in mc are about, but it works good enough to work with, at least under a X terminal. Mc seems to have some issues with col-width calculation when displaying filenames with unicode-characters. But that's more a design-issue than a functionality issue.
I'm in an UTF8 system now for 2 days. So far, no problems at all.

Thanks for the insight guys!

Did you make the change only in KDE, or did you also edit /etc/profile.d/lang.sh ?

I've never used midnight commander, so as long as 'man' works, I'm all ready to go

-y1

I did edit the /etc/profile.d/lang.sh.
I wanted to switch a while ago, but had to switch back as too many files had iso-content. Also some files had iso-characters in their filenames.
I had to correct these things for the final switch, which happened just days ago.

Good luck on your journey to UTF8!

Thanks again,

You raise an interesting point - what to do with all the files that have troublesome filenames... I've noticed a plethora of utf8-conversion scripts out there - do you have anyone in particular to recommend?
(While I do have backups, I'd rather not find out in retrospect that the particular script I choose is garbling all filenames beyond recognition)

-y1

Nope. I've written one myself but I won't recommend anyone to use it...
And I still have not converted everything. OOo-files, notes etc do still have some weird chars.
I basically relied on iconv and uconv.

I just used convmv and it did the trick perfectly of converting all my files to UTF-8. It has a "dryrun" mode and a live-ammo mode, and was even recommended by linux.com.

http://www.j3e.de/linux/convmv/

-y1

Basic utils still seem to lack UTF-8 support. For example will give just escape sequences for UTF-8 symbols.

Also, will just ignore UTF-8 characters, as if there were not any.

We are far away from good UTF support, it seems that many widespread GNU utils need a rewrite. If only I were a good programmist.. :)

Apart from the already mentioned mc vs utf8 issue, i can't saying anything to keep you from switching either. Configure your apps properly and you'll be happy ever after :)

Unfortunately it is not so. Sorry. Partly it is so. If you use US or British English or you don't use console at all then it will be so. But many issues are around, for examble those I mentioned above and plus those those packages that were updated in current, speaking of wget and oggenc they have the same problem: they span new lines while refreshing the lines.

You're right - Having searched savannah for info on findutils, I cannot find any open bugs on find not being utf8-compliant. Any ideas or pointers where to search for the root cause to this?

I mean - for groff/man, it's easy: lot's of useful patches sitting in cvs, but no release for ages (groff:4 years, man: 1 year) - but for findutils everything appears to be OK as per the bug database and changelogs...

Then again, I'm a novice at tracking these things down - hence the question :-)

-y1
(slowly returning online...)

The groff utf8 support is still broken. I cannot display any Chinese man pages. Please, slackware developers, either remove those Chinese man pages or add the groff-utf8 package:
http://www.haible.de/bruno/packages-groff-utf8.html