On 2003-03-19 at 12:06:08 [-0800], Axel D=F6rfler wrote: > Tyler Dauwalder <tyler@xxxxxxxxxxxxx> wrote: > > And while I'm thinking of it, Axel, can we add your modified=20 version > > of > > the iso9660 add-on to the cvs=3D3F Also, do you mind if I move the >=20 > Sure, but it will only work under R5 as of now :-) Why is that? > And please don't be to harsh on me - it's really some kind of a=20 hack=20 > ;-)) Hey, if it works, that's what matters for the time being. Thanks! :-) > > cpp.{h,cpp} files from BFS to somewhere more public so the rest of > > the > > kernel can share them=3D3F >=20 > Well, we can do it for now - if we'll find any problems with it at a > later date, we can simply remove them again. > The most public location for the header would be os/drivers/, a less > one private/kernel/.=20 Well, if there's no good reason not to have it in os/drivers, I guess=20 I'd vote for that. > The cpp.cpp should probably be part of the src/ > kernel/core directory, right? Sounds good to me. If they don't mysteriously move to those locations=20 by the time I get around to checking in my first round of udf stuff,=20 I'll go ahead make the move, okay?=20=20 > > > Similarly the other modules would get respective objects. With=20 the > > > Axel may for instance object, that C++ in the module interface=20 is > > > not > > > exactly something he likes -- which could be worked around by=20 > > > using > > > ugly ;-) C structures instead -- or something else I don't think > > > of, and > > > perhaps wouldn't have an answer to... > > Yeah, I like that idea. And C++ (or something very similar) works=20 > > for > > me. :-) >=20 > I am not completely against introducing C++ API in the kernel=20 itself - > but if we do, it *must* only be used for a specific case, and not be > usable by any other kernel components. Okay. > We shouldn't force anyone to switch to C++ to use a kernel service > right now. Why not? (I'm just curious, not challenging; I do enjoy making cranky=20 C programmers stretch their horizons ;-) -Tyler