Re: [Xen-devel] [PATCH] README: add Pixman as build dependency

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Tue, 2012-11-27 at 02:59 +0000, Matt Wilson wrote:
> On Mon, Nov 26, 2012 at 07:35:43PM +0000, Stefano Stabellini wrote:
> > On Mon, 26 Nov 2012, Roger Pau Monne wrote:
> > > On 23/11/12 13:51, Stefano Stabellini wrote:
> > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>
> > > > 
> > > > diff --git a/README b/README
> > > > index 21a81b2..88300df 100644
> > > > --- a/README
> > > > +++ b/README
> > > > @@ -51,6 +51,7 @@ provided by your OS distributor:
> > > >      * Development install of libaio (e.g. libaio-dev) version 0.3.107 
> > > > or
> > > >        greater. Set CONFIG_SYSTEM_LIBAIO in .config if this is not 
> > > > available.
> > > >      * Development install of GLib v2.0 (e.g. libglib2.0-dev)
> > > > +    * Development install of Pixman (e.g. libpixman-1-dev)
> > > 
> > > Should we add a configure test for Pixman? I guess this is needed by Qemu.
> > 
> > Yes, it is a QEMU dependency.
> > 
> > Is it actually a good idea to replicate all the configure checks done by
> > QEMU's configure script in the Xen configure script?
> 
> I don't think it's a good idea to duplicate checks.

Ack.

>  We should call
> qemu's ./configure early, perhaps from our ./configure, and

Problem is that tools/qemu-blah/configure isn't available until it is
cloned, which happens at build rather than configure time.

>  make sure
> that errors are clearly propagated and messaged to the user.

We should strive for this no matter when/where the qemu configure is
run. So long as qemu's configure fails cleanly and we abort when that
fails we should be in reasonable shape?

Ian.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel