Re: possible bug in libtool w/ MinGW on Chinese locale

[Pete Batard]

Hi Ralph,

Thanks for getting back to us.

On 2010.03.14 07:47, Xiaofan Chen wrote:
> On Sat, Mar 13, 2010 at 6:31 PM, Ralf Wildenhues<address@hidden>  wrote:
> > 1) Would it also be sufficient to just
> >   LC_ALL=C; LANGUAGE=C; export LANGUAGE LC_ALL
>
> This is enough to solve the issue.

Yes, I don't think we need to set all the variables. Our only concern 
here is having to require our users to set anything manually before they 
can run make, otherwise, any set of variables that works is fine.

> > or do you actually need LANG to be set?

From Xiaofan's tests, we probably don't need LANG to be set then. But 
definitely need some of these locales set to prevent the libtool issue. 
Was LANG the only problematic variable for SCO?

> > Autoconf-generated configure
> > scripts set these two unconditionally, so I don't expect any problems
> > there.
>
> Somehow under MinGW I had the issue even though the configure
> scripts are generated by Autoconf.
> http://old.nabble.com/Re%3A-Problem-compiling-with-MingW-p27809936.html
> With the forced Locale settings and it became fine.
> http://old.nabble.com/Re%3A-Problem-compiling-with-MingW-p27810019.html

Indeed, it seems we can only get libtool working, on a (some) systems 
where the locale variables are not set, unless the libtool script sets 
some of them.

> > Tim, this code was from some initial revision in '97.  Do I gather
> > correctly that it has no relevance today, and that we should just
> > remove this?
>
> Pete's response here. We think that SCO may not be relevant now.
> http://old.nabble.com/Re%3A-Problem-compiling-with-MingW-p27816064.html

If the SCO specifics are no longer relevant, forcing the locale 
variables unconditionally in future versions of libtool would be our 
preferred solution, but we obviously don't know if that's a possibility 
for you.

Regards,

/Pete