Re: [PATCH 0/5] Refactoring to avoid namespace collision on AIX

On Thu, Jun 18, 2015 at 12:57:38PM -0700, Bryce Harrington wrote:
> On Wed, Jun 17, 2015 at 10:31:47AM +0100, Chris Wilson wrote:
> > On Wed, Jun 17, 2015 at 02:05:35AM -0700, Bryce Harrington wrote:
> > > This patchset initiated as a fix for bug #89339.  The actual fix is the
> > > last patch in the patchset.  Some of the proposed fixes really just
> > > worked around the actual problem, yet none-the-less were worthwhile
> > > refactorings that would cleanup the code, so are submitted along with
> > > the fix.
> > > 
> > > Bryce Harrington (5):
> > >   cairo-script: Always include config.h first thing
> > >   cairo-script: Add missing copyright and boilerplate
> > >   cairo-script: Cleanup boilerplate header for consistency
> > >   cairo-script: Prefer cairo from local tree
> > >   cairo-script: Rename struct member to avoid name collision on AIX
> > Reviewed-by: Chris Wilson <[email protected]>
> > 
> > Note that we have used #ifdef HAVE_CONFIG_H universally, but since we
> > control the build system we should probably depend on it like 1/5 here.
> 
> Do you mean that in other places in the cairo codebase we should be
> depending on config.h directly (without the cpp conditional)?  Or just
> here in the util module?

We should be consistent. I don't know of any machine that must provide
the defines on the compile line and never generates a config.h, nor can
I imagine someone willingly doing that to themselves. (Rather than
modify config.h, rewrite the CFLAGS!)

So I am in the mood to lose all the #ifdef HAVE_CONFIG_H
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
-- 
cairo mailing list
[email protected]
http://lists.cairographics.org/mailman/listinfo/cairo