"Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx> wrote: > "Ingo Weinhold" <bonefish@xxxxxxxxxxxxxxx> wrote: > [...] > > The second newsletter article Charlie was referring to is a good > > introduction (does anybody know, where to get the sample code?). > > I'd > > like to have more details, though. I've been trying for months to > > convince Axel to write an article, but he's just too lazy. ;-) > > Yes, I am :-) > You don't have to convince me to write one, because I wanted to do > that > anway, it's just that I need to find the time to do so :-) Uh, I've always considered that a lame excuse of yours. ;-) > > You might also want to have a look at src/add-ons/kernel/ > > file_systems > > / > > bfs/cpp.{h,cpp}. > > I think, cpp.cpp should be linked into the kernel and cpp.h made > > available as public kernel header. > > As long as we don't create a dependence on a specific compiler > version, > I agree with you here. OTOH that step would require a newsletter ;-) :-) [...] > > > And here are a couple I'm specifically curious about: > > > - new and delete > > No problem. The operators have to be overridden as in cpp.h. I > > dislike > > the `#define new new (_dontthrow)' there, though. And I don't think > > `extern nothrow_t _dontthrow;' is needed, for `nothrow' should be > > defined anyway. > > Yes, you are right "nothrow" is already defined and works okay (will > update it shortly). > What would you suggest to replace the #define with? Remove it completely, I would say. As the brave userland coders do, one can use the nothrow variant explicitly. Or override the non-nothrow versions as well -- I'm not even sure, if that is necessary, since we compile with `-fno-exceptions' anyway. CU, Ingo