On Mon, Mar 15, 2010 at 11:03:09AM -0500, Nicolas Williams wrote:
> On Mon, Mar 15, 2010 at 09:53:04AM +0100, Marcel Telka wrote:
> > On Sun, Mar 14, 2010 at 07:32:34PM -0500, Nicolas Williams wrote:
> > > > >  - Would it be better to patch configure.ac instead of configure?
> > > > 
> > > > When I would patch configure.ac the configure needs to be either 
> > > > patched too or
> > > > regenerated. Regeneration on the fly needs some extra tools like 
> > > > autoconf.
> > > > Patching of both configure.ac and configure would technically end in 
> > > > very
> > > > similar situation as I have now where I patched the configure only. 
> > > > During the
> > > > build nobody cares about the configure.ac content.
> > > 
> > > Huh?  We don't have autoconf installed on build servers?
> > 
> > We do. But additional complexity is not good if I can easily avoid it. :-)
> 
> Ok, though you might want to ask the SFWNV GKs about that.  Personally I
> think we should patch configure.ac, run autoconf, then ./configure, but

In general, yes.

> don't feel strongly enough about the matter.  (My preference results
> from: a) it's easier to understand and patch configur.ac than configure,
> b) that's what you'd have to do in order to contribute patches upstream.
> In this case the patch is already in the head upstream, so there's no
> harm in patching only ./configure.)

I agree. The current configure patch is just an interim solution until new mutt
is released upstream.


Thanks.

-- 
Marcel Telka
RPE, Systems

Reply via email to