[arch-dev-public] replace man with man-db?

Aaron Griffin aaronmgriffin at gmail.com
Mon Dec 22 13:54:41 EST 2008

Reply from the man-db upstream maintainer:

---------- Forwarded message ----------
From: Colin Watson <cjwatson at debian.org>
Date: Mon, Dec 22, 2008 at 12:30 PM
Subject: Re: [arch-dev-public] replace man with man-db?
To: Aaron Griffin <aaronmgriffin at gmail.com>
Cc: arch-dev-public at archlinux.org

On Mon, Dec 22, 2008 at 12:04:47PM -0600, Aaron Griffin wrote:
> On Mon, Dec 22, 2008 at 11:57 AM, Andreas Radke <a.radke at arcor.de> wrote:
> > This was brought up a while ago:
> > http://bugs.archlinux.org/task/9130
> >
> > There's a usable man-db pkg in AUR that could replace man in core after
> > some testing.
> >
> > I haven't heard of many issues with localized man-pages. Reading around
> > Gentoo and Fedora still use "man" while Debian only provides "man-db".
> > There are many patches in the Fedora cvs to solve the claimed issues
> > with man. We have not one bug report for our "man" package.
> We do have a few localization bugs with regard to... Russian or
> Ukrainian man pages, or something...

That's not surprising, as they typically use ISO-8859-2 and getting
groff to deal with anything other than ISO-8859-1 is hard work. This is
the sort of area where trying to fix up man will really bite you over
time, as it simply doesn't have enough of the logic you need to deal
with encodings properly (you can sort of bodge it as long as there's
only ever one encoding in use for a given language, I think, but that's
a totally unsound assumption in the real world). I put a lot of effort
into getting this right in man-db; I'll fix any such remaining problems
in man-db promptly if they're reported to me.

Do you use the Debian multibyte patch to groff? I concede that it's
pretty gross and we've had problems getting it updated to the latest
upstream version of groff, but it does work for this sort of thing (it
provides an 'ascii8' device which is not entirely typographically sound
but which passes through single-byte data in the same encoding). If not,
you would probably need to use groff from CVS rather than the latest
release, since that provides an infrastructure for converting manual
pages into a neutral encoding that can be handled by groff (but it
doesn't yet have enough intelligence to get things like Japanese
"kinsoku shori" line-wrapping right).


Colin Watson, man-db upstream maintainer           [cjwatson at debian.org]

More information about the arch-dev-public mailing list