Re: master b1e9151: Enable the native display of BMP images on Haiku

[Eli Zaretskii <eliz@gnu.org>]

> From: Po Lu <luangruo@yahoo.com>
> Cc: emacs-devel@gnu.org
> Date: Sat, 26 Mar 2022 21:47:50 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> From: Po Lu <luangruo@yahoo.com>
> >> Cc: emacs-devel@gnu.org
> >> Date: Sat, 26 Mar 2022 21:22:41 +0800
> >> 
> >> Eli Zaretskii <eliz@gnu.org> writes:
> >> 
> >> > It's better not, I think.  We don't do that elsewhere, do we?
> >> 
> >> I think there are a few cases, though probably not deliberate.  If you
> >> insist on removing the definition in haikuimage.c, I don't see why not.
> >
> > What are the arguments for leaving it in haikuimage.c?
> 
> Mainly so that if the DEFSYM in image.c is moved under some different
> conditionals (or removed entirely), haikuimage.c will continue to work.

Then I don't think we should keep it in haikuimage.c.  If it is ever
removed, haikuimage.c will fail to compile, and the error message will
tell exactly what's wrong.  We have similar issues with symbols used
in w32font.c that are declared elsewhere, for example.  Why would
haikuimage.c be treated differently?