Conformance tests

[Austin English]

On Fri, Aug 14, 2009 at 3:10 PM, Keith
Muir<keith_kw_muir at blueyonder.co.uk> wrote:
> What disturbs me about the conformance tests is that yes there are 0%
> failures but when you check you discover its 0% failures on virtual
> machines. Failures on real hardware under XP can reach as high as 6% with
> peaks greater than that. Windows is tolerant of all types of hardware wine
> still isn't. I may be stating the obvious but hardware tolerance must be a
> goal for the wine project if it isn't already.

Those conformance tests shouldn't care about hardware (graphics tests
aside). When testing 'bare iron', you're likely to be testing
non-clean installs. Virtual machines have the benefit that they can be
reset each day, so e.g., old cruft from tests, other applications,
etc. aren't there to mess up the results.

A lot of the 'real' machines also aren't necessarily english locales,
or have a different DPI set, which breaks some tests.

-- 
-Austin