pciaccess conversion the revenge..

[Daniel Stone]

On Tue, Mar 11, 2008 at 06:32:49PM -0300, Paulo Cesar Pereira de Andrade wrote:
> Daniel Stone wrote:
>> On Tue, Mar 11, 2008 at 04:24:13PM -0300, Paulo Cesar Pereira de Andrade wrote:
>>>  Not really, if the developer cannot test the changes.
>>
>> If you can't test the changes, then you can't test the changes, full
>> stop: no matter whether you're changing the driver or the server.
>
>  Actually, writing a (temporary) compatibility module would fix
> problems without need of testing, or recompile, assuming it would
> export the same abi and functionality, and enough tests are done
> in the "available" hardware.

That's a big assumption.  Assuming you write perfect code and manage to
get all the PCI semantics correct.

>  Anyway, if the X Server and modules keep changing api/abi, some
> method, tinderbox or whatever must be enforced. So that developers
> will at least know things are broken, and don't "carry" either
> unbuildable or "plain broken" system for so long.

http://tinderbox.x.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.x.org/archives/xorg/attachments/20080312/b432b036/attachment.pgp>