LANG=fr_FR.utf8 vs. french man pages
 

Hi,

I'd like to have my Slackware 14.0 system completely in French, including man pages. I have a little problem with these.

Here's what I put in my /etc/profile.d/lang.sh:

And here's /etc/rc.d/rc.font:

A stock Slackware 14.0 system already ships a handful of man pages translated into French, under /usr/man/fr. For example :

Or:

Unfortunately, these man pages display the usual suspects of random character soup instead of french special characters like "é", "è", "à", "ç", and so on.

I have the same problem when I install some external (UTF-8-encoded) french man pages from here: http://manpagesfr.free.fr. So I vaguely guess the problem doesn't originate from the pages themselves.

Hmmm. As is often the case with problems related to encoding, I'm clueless.

Hi,

As usual, there are only awkward solutions. You can edit your /usr/lib/man.conf to force groff to interpret the man-pages as an UTF-8 encoded source (the French man-pages are rendered broken because groff sees them as latin1-encoded):

But then, the man-page encoded in latin(9|1) will be displayed broken. The second solution is to convert the French man-pages into iso-8859-1 using iconv. That's what I used to do when I was installing them (they're not up-to-date anymore, be careful). The conversion works well, except for the pages about the different charsets in 7, but who reads that?

Thanks very much! I fiddled around some more and finally managed to get it right. I converted all the man pages to latin1 before installing them, and they display OK now.

Cheers,

Niki

I also use "LANG=fr_FR.UTF-8".

With Slackware (and maybe with other distro), it seems that french man pages are shipped in various encodings:

man pages located in the fr.UTF-8 sub-directory are correctly displayed. For example:

Other pages not stored in this sub-directory won't display correctly most probably because they are not UTF-8 encoded.

So I suspect that you just need additional french man pages encoded in UTF-8 with no need to change any man configuration options.

Nope. :)

vim.1 is correctly displayed because encoded in latin1 and therefore rightly converted in UTF-8 (� means "non-sense in UTF-8"):

/usr/man/<local>.<charset> are directories for man-db, not the implementation of man Slackware uses:

NonNonBa is right. Manpages have to be latin1-encoded to be displayed correctly with french special characters.

OK, I was wrong: I was just trying to advance the topic :)

I'm interested in knowing how you did the conversion to latin1 of existing man pages (ex: nmap). Thank you.

I didn't. Encoding for this one is still wrong.

Guess I'll have to write a little script that scans /usr/man/fr for UTF-8-encoded pages and automatically converts them to LATIN1.

For the French man-pages archive, there's an option in the configure to perform the conversion. I don't remember if the maintainer had included the other build process I made for man-pages-extras.

For the broken pages of the other packages, there's another solution I haven't thought to:

Thus, you don't need to touch anything. If a page is rendered broken, just reload it with uman.

This is a GREAT idea. You deserve at least a dozen "found this helpful" clicks for this one.

Thanks a lot!

IMO the solution proposed by NonNonBa is of course the best one, at least until all man pages in Slackware be shipped in UTF-8 encoding.

@NonNonBa aka Seb: By the way, do you remember your promise to launch the "UTF-8 Slackware" project? ;)

Thanks for the tip
 

Thank you. I'm glad I learned something from my mistake :)

@kikinovak, gegechris99: You're welcome. :hattip:

@Didier Spaier: Yes, I remember I promised a thread. ;)
I wanted to finish the howto about TeX Live first, and caught a nasty little virus in the meanwhile. But I didn't forget, it's in the TODO, don't worry. :)

Of course the cleanest solution would be a script that scans everything in /usr/man, hunts down utf8-encoded man pages and converts them to latin1. This is slightly over my scripting abilities, but I've started to experiment nevertheless.

I would certainly prefer the other way round. UTF-8 is the future.