RFC: detecting wine drivers in the audio tests
Robert Reif
reif at earthlink.net
Wed Apr 30 06:45:49 CDT 2008
Maarten Lankhorst wrote:
> Hello Robert,
>
> 2008/4/29 Robert Reif <reif at earthlink.net>:
>
>> The returned result of some audio functions on windows may be inconsistent
>> because a driver may actually supply the returned value.
>>
>> This presents a problem for the wine regression tests because a buggy
>> driver may return an unexpected result which causes the test to fail. One
>> way around this is to accept known failures as OK but that reduces the
>> usefulness of the test for wine because it may allow wine bugs to slip in.
>> I'm proposing the we determine if we are running on wine by defining a wine
>> manufactures id and checking for that id in the test. If a wine driver is
>> found, don't accept a failure as OK. The returned result should be well
>> defined and any failure is unacceptable. However, if the driver is not a
>> wine driver and there are known buggy windows drivers that return specific
>> errors, we can check for that and not fail the test. This should be easier
>> than maintaining a database of known broken drivers and black listing them.
>>
>> I am concerned that requiring 100% wine regression test success on windows
>> is not practical when there are broken windows drivers out there. Accepting
>> the failures as success is not good for wine because it may allow buggy wine
>> drivers to also pass. I think we should hold wine audio drivers to higher
>> standards than the typical audio card manufacture.
>>
>> I am suppling a minimal patch to the alsa driver and a single wave test to
>> illustrate this concept. I hope this allows valid tests to remain in spite
>> of buggy windows drivers.
>>
> I don't know, the general idea is that whatever windows lets you do is
> fine for wine, so I'm not 100% sure that testing for this is a good
> idea, because it means in border cases windows programs can't rely on
> those to return a correct or incorrect value either.
>
>
The point is to not throw out a valid test because on in a million PCs
has a buggy sound driver. I doubt that any programs would rely on a bug
being present to work properly. Particularly ignoring a specific
failure code.
> By the way you also used a C++ style comment, wine doesn't allow that.
>
>
This is just an RFC. The example code is not ready for acceptance.
Doing this right will require careful auditing of the test and I want to
make sure this is reasonable before proceeding
> Cheers,
> Maarten
>
>
>
>
More information about the wine-devel
mailing list