On December 6, 2002 03:12 pm, Alexandre Julliard wrote:
I don't know, it feels cleaner to have a separate wrapper; I'd suggest doing at least the first version that way, and once things are cleared up maybe we can find a way to merge everything into winebuild cleanly.
Yes, I agree it feels cleaner to keep it as a wrapper, but current practice points in the other direction. Nevertheless, I fully agree that it makes sense to have the first cut done separately, so we better understand what the problem is before we do anything else.
That being said, if we do plan to to integrate it into winebuild, I think we should do it sooner, rather than later, even at the expense of small temporary ugliness (e.g. fork winebuild from itself) in winebuild. We are in Wine 0.9 mode right now, and we should try hard to stabilize the utilities, and the *process*. The importance of said stabilization can not be overstated, as a bunch of other things depend on it, such as documentation, and other scripts (such as winemaker). And these dependents have been historically a _lot_ slower to change than C code. But most importantly, if people start using Winelib, it's going to be very hard to change established process as everyone's build system depends on it.
Anyway, I will try to finish winewrap in it's current form, and we can evaluate merging it into winebuild when I'm done.