It seems that every man page written in french that uses the iso-8859-15 caracter set is not correctly encoded. Reproducible: Always Steps to Reproduce: 1. emerge man-page-fr-1.64.0 2. man iso-8859-15 3. Actual Results: Le jeu de caractA"res ISO 8859-15 en octal, dA(C)cimal et hexadA(C)cimal. Expected Results: Le jeu de caract
It seems that every man page written in french that uses the iso-8859-15 caracter set is not correctly encoded. Reproducible: Always Steps to Reproduce: 1. emerge man-page-fr-1.64.0 2. man iso-8859-15 3. Actual Results: Le jeu de caractA"res ISO 8859-15 en octal, dA(C)cimal et hexadA(C)cimal. Expected Results: Le jeu de caractères ISO 8859-15 en octal, décimal et hexadécimal. This happened after emerging the latest version (1.64)
Hmm, these manpages need conversion to utf-8 anyway (at least w/ unicode flags set).
Same as here. The previous version works great !! why did you erased it ? PLEASE STOP ERASING paquets from portage when they work !! Specially when the new version is buggy !!! Do T have to quickpkg before every update ??? I think it is not a good way to maintain a "stable" branch...
You can see it raw in some pages for exemple man lp is working: the
You can see it raw in some pages for exemple man lp is working: the é char is coded E9 and man dd is not working: the same char is coded C3 A9 i don't know more about it.
same here !!!!!!!!!!!!!
(In reply to comment #4) > same here !!!!!!!!!!!!! Sure, and what's the point of making noise here?
9 months old and still not fixed ? :\
(In reply to comment #8) > 9 months old and still not fixed ? :\ > vapier how the hell do I fix this?
dont read french man pages there is no such thing as "wrong encoding" with man pages ... there is a desync with handling of unicode in most man pages best to sort it out with man itself rather than screwing with man-page packages themselves, but i dont really delve into this crap so i'm just guessing
> dont read french man pages my students pretty much have to, as their command of english is not that great. > there is no such thing as "wrong encoding" with man pages that statement is impractical since, as there is no absolutely reliable way to guess the encoding of a file, a robust installation of man pages ought to ensure that they are all using the same encoding. The ebuild could use recode to ensure this. in the mean time, I have been using a heuristic filter for man. In /etc/man.conf I have: NROFF /home/denys/bin/to-latin1 | /usr/bin/nroff -Tutf8 -mlatin1 -c -mandoc and to-latin1 is the following python script: #! /usr/bin/python import sys, chardet, subprocess, os text = sys.stdin.read() enc = chardet.detect(text)['encoding'] pro = subprocess.Popen(("/usr/bin/recode", "%s..Latin-1" % enc), stdin=subprocess.PIPE) pro.stdin.write(text) pro.stdin.flush() pro.stdin.close() os.pidwait(pro.pid)
> os.pidwait(pro.pid) oops! that should be: os.waitpid(pro.pid, 0)
As written here : http://manpagesfr.free.fr/ All versions of man-pages-fr should be UTF-8 encoded since 1.59.0. A rapid test on the 2.39 package file gives only utf-8 (some pure ascii) # for i in $(equery f man-pages-fr | grep "bz2$"); do file -iz $i; done Using utf8 locale such fr_FR.UTF-8, we should have in /etc/man.conf : NROFF iconv -t utf-16LE | /usr/bin/groff -Tutf8 -mandoc conversion to utf-16LE, is to be able to have -Tutf8 option to groff without having double utf8 conversion or desctructive/in error conversions. If one is in an non-utf-8 locale, one could just add iconv conversion to it: NROFF iconv -t utf-16LE | /usr/bin/groff -Tutf8 -mandoc | iconv -t iso8859-15 maybe do this depending on a unicode use flag ? Btw, man-pages-fr are now 2.79 and we have man-pages-extra-fr, always according to http://manpagesfr.free.fr/. I think the link for this package should point to http://manpagesfr.free.fr/ instead of http://fr.tldp.org/manfr.php which has not been updated since 2003 ! I may open some new bugs for the last two remarks if needed.
Closing bug about obsolete, no longer shipped version.
