[RFA/commit 2/2] Import gnulib's errno module.

[Eli Zaretskii]

> From: Tom Tromey <tromey@redhat.com>
> Cc: brobecker@adacore.com, gdb-patches@sourceware.org
> Date: Tue, 18 Dec 2012 14:30:11 -0700
> 
> >>>>> "Eli" == Eli Zaretskii <eliz@gnu.org> writes:
> 
> Eli> How about
> Eli> 3. Test for either ENOENT or EILSEQ where we now test for EILSEQ.
> Eli> ?
> 
> It is tempting but I think it exposes us to libiconv's internals too
> much.  That is, if libiconv changes its choice here, and we are using
> the gnulib EILSEQ, then gdb would start doing the wrong thing.

How about limiting that to known versions of libiconv, then?

I mean, I agree that this all is kludgey, but the whole issue is a
mess.  A library should not define a global symbol, it should only
define symbols for its own private use.