Re: login.conf --> UTF-8

On Wed, Apr 02, 2014 at 09:31:08PM -0400, Daniel Eischen wrote:
> On Thu, 3 Apr 2014, Steve Wills wrote:
> 
> > On Wed, Apr 02, 2014 at 03:56:35PM -0700, Sean Bruno wrote:
> >> On Wed, 2014-04-02 at 18:06 -0400, Garrett Wollman wrote:
> >>> In article <1396457629.2280.2.camel_at_powernoodle.corp.yahoo.com>,
> >>> sbruno_at_freebsd.org writes:
> >>>
> >>>> I'd like to make this change to login.conf for default installs.
> >>>>
> >>>> This removes some amount of hackery in the ports system that is working
> >>>> around our lack of UTF-8 in the base.
> >>>
> >>> I'm not sure what the connection is here.  Surely the ports system
> >>> runs with the locale of the user running "make" (which in my case is
> >>> going to be "C").  Any port that requires a specific locale to build
> >>> properly needs to be setting that locale explicitly.
> >>>
> >
> > You'd think so, but that's not what's happening. What's happening is the
> > software builds as long as the locale isn't C. Hence, ugly hacks like this:
> >
> > http://svnweb.freebsd.org/ports/head/Mk/bsd.ruby.mk?annotate=348863#l257
> >
> > Why? Because the people writing it have never encountered a system where LANG
> > isn't set or is set to C. Yes, it's a bug in their software. No, they never
> > have and never will encounter it. Because every other operating system sets
> > LANG to whatever the user specifies. And so they have no interest in fixing it,
> > because neither they nor any one they know will ever encounter it, and even if
> > you report it to them they will tell you it's a bug in your system for not
> > having LANG specified. And I have no interest in patching it hundreds of
> > times.
> >
> > And this is just one example. There are others, I think, that aren't ruby
> > related at all.
> 
> The first thing I do when I get a Linux system is set LANG to C.
> I hate all the colorizations and incorrect ordering from ls when
> LANG isn't C.  So you are saying, that ports will be broken when
> I set LANG back to C again?
> 
> >> I have been informed by folks that this change I suggest would help in
> >> the case of ports having to declare UTF-8 support explicitly or
> >> something. I'm hand-wavy on the details and ignorant of the hacks in
> >> place.  I only know that I've been *told* this.
> >
> > I think we should join the club of asking the user, but that's more work and
> > until then having a reasonable default and having people change it seems sane.
> 
> A default is fine, but saying that ports will be broken when not
> using the default is not fine.  This is LANG, not a gcc/clang
> machine-specific optimization that someone has set to get an
> extra 0.001% improvement, but happens to break the compiler for
> some ports.

I suppose you're right. Ugly hacks to work around ugly hacks will stay. :)

(Not that I'd planned to remove them any time soon anyway, because such a
change would take a long time to propogate to all supported versions anyway.)

Steve