Am I missing something here?
Not sure. I was under the impression that Winelib apps used wineserver, the protocol for which won't be frozen until 1.0 - this might be what he meant.
On Thu, 2002-11-21 at 14:51, Shachar Shemesh wrote:
Steven Edwards wrote:
The problem we have with people making and shipping winelib apps at this point is that with every release of WINE untill 1.x we will have breakages.
If winelib contains what I think it contains, I don't see why that should happen. Are you sure about that point?
After all, a winelib app is a linux/unix app that is dynamically linked (by name, I'm sure) with winelib dlls that are, today, part of wine. Both the functions linked and their parameters are defined by MS and MSDN, and will not change, ever. Bugs may be solved, but as the application was originally designed to work with MS's implemnetation, and as a bug is defined to be "incompatible with MS's implementation", I don't think it possible for a winelib app to rely on a wine bug for functionality. I therefor don't see how a breakage can happen unless we have regression in wine. As such, breakegaes should be very rare, not often. 1.0 or not.
Am I missing something here?
If you want to help people that produce OSS windows applications support WINElib then for now, at least untill 1.0 we can only support building winelib apps and not binary winelib applications.
That point is 100% understood assuming I accept your previous statement (which I don't). Reply only if you think this statement holds true even if your previous one doesn't.
Thanks Steven