Wine 0.9 progress

Vincent Béron vberon at mecano.gme.usherb.ca
Fri Aug 29 19:23:47 CDT 2003


Le ven 29/08/2003 à 20:13, Dustin Navea a écrit :
> so without the xp server flavors and all the 2k3 flavors, we have 16..  even
> if we simplify to nt4, and 2k, its still a dozen like you originally said,
> although if you take out the osr's then it would be 8, but there are a couple
> of service packs and osr's and things of that nature that im sure changed
> some things that would affect the conformance test suite..  either way, we
> need to have 16 people for at least the first run (or a few people with
> multiple os's on their comps) to find out which service packs and which
> flavours of the nt based windows' actually have changed things, to where we
> would need to continue to run a test on those specific service packs, etc.. 
> confused yet? 

I follow. The problem with this is we'll need the 16 for longer than
only the initial run, as not all API have a conformance test yet. If
it's not tested, we don't know if it changed or not.

> 
> say nt4's sp3 and sp6 changed some stuff that would affect the results of the
> conformance tests.  we would do the first run on all nt service packs
> (finding them all would be another story, though).  then we would discover
> that nt4 reg, sp3, and sp6 gave different results on the tests, so we would
> run subsequent tests on only those 3, thereby eliminating about 7 other tests
> (sp1, 2, 3a, 4, 5, 6a, etc)

Overkill. Wine advertises NT4 SP6 only, not the other SPs. Same thing
for each major version, only one SP match. So there's little point in
conforming to something not supported by Wine, as it's been superceded.
Future SPs will be tested as a "new" version, and if conformance tests
fails on them, then it'll be fixing time (conformance test, plus Wine if
applicable).

Vincent





More information about the wine-devel mailing list