[Note: This e-mail get's into some additional design discussions, so I'm
adding anybody who has responded to the DRM discussion to the CC list,
to circumvent the sourceforge e-mail log jam.]
Alan Hourihane wrote:
>
> On Fri, Apr 05, 2002 at 07:46:59 -0700, Jens Owen wrote:
> > > (II) I810(0): [dri] Buffer map : ca3000
> > > (II) I810(0): [drm] added 256 4096 byte DMS buffers
> > >
> > > Fatal server error:
> > > Caught signal 4. Server aborting
> >
> > I'm working on a better means of reporting this error.
> >
> Jens,
>
> If I understand this correctly......you could use something like this...
>
> if (!xf86LoaderCheckSymbol("drmCommand"))
> ErrorF("OOPS you need to grab a later libdrm.a\n");
Alan,
I think your suggestion would be the easiest to implement and deploy at
the individual driver level. However, I'm concerned that we're not
using the revision system we've built into the DRI (when the full screen
extension was added) and that using piecemeal checks like this gets
confusing when you get many of them together.
I recommend that we bump the DRI revision number to 4.3.0 and update
[radeon, i810, mga, gamma, r128]_dri.c to check for this version. We
then add your symbol check to dri.c.
The side effect is that users will also need to update libdri.a as well,
but we can package both libdri.a and libdrm.a together. Doing this work
will break some ground on how to update new extensions that are used by
all drivers (like the full screen extension).
-- /\
Jens Owen / \/\ _
[EMAIL PROTECTED] / \ \ \ Steamboat Springs, Colorado
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
