On December 3, 2003 10:57 am, Vizzini wrote:
We're still working out a plan for sharing headers; unfortunately it's not as easy as just using Wine's headers.
I know things are not always as easy as they seem, and I must apologize for reopening this issue, but there's something I miss. Namely, we're aiming to be MS compatible, so in theory it _should_ be as simple as just using Wine's headers.
I am aware we're not there yet, but I think we're doing OK. Most of the ifdefs in our header files for __WINESRC__ are to restrict the headers. That is, we simply don't allow some things to be used in our source. This should be no problem for you, no? If there are incompatibilities, we are very interested in fixing. In fact, the Winelib porting effort I started a year ago is aimed (in part) at finding problems with our headers.
Also, Steven has tried to multiple times to explain that you guys felt the need to build your own headers because ours are not perfect. I am sorry to say, this makes no sense to me. First, as you can see, it's not the case that if you're building them yourself, they are going to be perfect. Second, you're duplicating an enormous amount of effort that could be expanded in helping us fix our headers. Third, it's a lot better to work off of something (that's not that bad to begin with), then to roll your own (availability, time-to-market, duplication of effort, maintenance, to name a few important reasons).
Just for my edification, what would happen to your tree (or what chances would entail) if you just switched to our headers?