So, we've got our first bug on the ForceGallium option patch¹, where a
user with self-compiled mesa gets bad results.  Which annoyed upstream
when the user (quite reasonably) reported a bug upstream.

As I see it, there are two possible fixes for this:
1) Better documentation: ForceGallium could print some warning text in
Xorg.0.log and the man page could mention this interaction.

This has the advantage of being (reasonably) easy to globally switch
between classic and gallium drivers.  Also, the work's already done :).

2) Better patch: it would be quite easy to ship the fallback driver
in /usr/lib/dri-alternates and add this to the mesa dri searchpath.
This covers direct rendering; AIGLX requires a reasonably easy X patch
to know how to fallback.

This has the advantage of being potentially upstreamable, more general
(although there probably won't be another driver we need to do this
for), easier for users to test individual apps with different drivers
(using LIBGL_DRIVERS_PATH to switch), and doesn't break non-ubuntu
mesas.

I'm not really sure which of the tradeoffs to give priority to here,
hence the call for comments.

[1]:
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/715939

Attachment: signature.asc
Description: This is a digitally signed message part

-- 
Ubuntu-x mailing list
[email protected]
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-x

Reply via email to