Re: [gentoo-dev] Re: Disabling locale at emerge output

[Jeroen Roovers <jer@gentoo.org>]

On Wed, 27 Apr 2011 02:47:46 -0600
Ryan Hill <dirtyepic@gentoo.org> wrote:

> > Are you saying you are setting up the tree-wide project that
> > automatically translates LC_MESSAGES? Round of applause!
> 
> Nope, but I am raising donations to fund an instructive manual on the
> usage of the NEEDINFO resolution.

As already explained, requesting information and temporarily closing
bug reports happens a lot already, but usually greatly stalls a
bug's resolution. Developers have to make time to investigate the
problem and lose time when they have to wait for additional information
and need to reschedule.

> > > For a growing portion of our user base it's not "nice" to see
> > > these messages in their native language, it's fucking "necessary".
> > 
> > Another round of applause for the appalling attitude.
> > 
> > Could you elaborate on the need to read error messages in your own
> > language? How is the sys-apps/portage translation project doing
> > anyway? Still only in Polish, hm? :)
> 
> Que?  Are you saying that because portage is in English that people
> that don't speak English don't use Gentoo?  I would have to disagree.

Obvious straw man. 

> If you think that getting the occasional bug report in a language you
> aren't familiar with is annoying,

You are obviously unaware of the scale of the problem. Devoting time to
a bug becomes a serious problem when requested information takes a day
or more to arrive, if at all.

> imagine how not being able to get any build output you can read
> might feel, especially if the reason is only because some dev
> couldn't be bothered to occasionally ask someone to repost an error
> message in English.

You underestimate the vastness of my imagination.

> This reminds me of the time someone wanted to ban inline `emerge
> --info` postings in bugzilla (ie. attachments only) because they
> found scrolling through them bothersome.

I think you mean the time when I was requesting ATs not to post such
information on stabilisation bugs just to confirm it was alright to
stabilise. You know what? - I'll talk on that obviously tangential
argument. That (arch team specific) policy was apparently abandoned,
because it doesn't happen now. Perhaps I didn't cry victory on that one
--- perhaps I should have specially informed you, since you seem to
think this reflects on any future argument. A course in informal
logic might do you  well.

You still haven't elaborated on the need to read build output in
non-English, btw. You could have argued for non-English users' rights,
or pushed for that LC_MESSAGES auto-translation project (I wasn't
kidding), or you could have argued for a bug reporting tool that
includes all the right information in the right language automatically.

Look here, I am even willing to argue on your side just to possibly
extract a meaningful objection to fixing LC_MESSAGES in
sys-apps/portage.

I can explain my position better on having LC_MESSAGES in the same
language as sys-apps/portage appears to non-Polish readers, though.

When someone has to read the English error output that sys-apps/portage
tacks onto the end of the real error (possibly output long gone from
the scroll buffer because it was an early sub-make that failed),
doesn't want to go all technical and find that error message, and
therefore needs to be asked to attach the entire thing or does that of
her own accord, then why should the actual error message
("foo/include/header.h: No such file or directory", in Simplified
Chinese) be printed in a language the developer cannot read?

The user only cares about using blender or openoffice - the user
couldn't care less about the crufty build system of a DEPEND packages
needed only at build time, or the language it prints.


     jer