Open MSch opened 13 years ago
ad 2. NK_INLINE
doesn't really fall under optimization, but rather under consistency with Apple frameworks.
Is NS_INLINE
available in our headers? We should use that.
ad 2. I meant your comment about future LLVM version. If that's the reason for NK_INLINE
then NK_INLINE
is premature optimization
"Be friendly to people that want to copy and paste stuff into their own source" I like that idea. We shouldn't depend to heavily on meta-language. NK_INLINE is probably something we don't need. But I'm sure there will be macros that are worth writing, and we have to decide between ultimative modularity or a more efficient (less LOC) framework, i decide for the later. We thus should decide that on a case-on-case basis, and I trust you guys to make that decision.
I just think NK_INLINE is something not really heavily used - but please, correct me if I'm wrong.
On the other hand, Erik introduced me to the great shorthand "_" instead of NSLocalizedString(@"", @"") - and I'm all for using that, even in categories (that may just include NKShorthands.h). If I really wanna use just parts of the framework, I can
a) add the macro file with it (it's just preprocessor stuff, it doesn't hurt or exists globally) b) add the one line to the .m header c) use search/replace
Don't forget guys, we're building a better tool for ourselves. Other, imaginary third people should be considered, but our needs come first.
Here's how I understand FoundationKit (as opposed to PSFoundation)
@eaigner added the
NK_INLINE
macro to NKShorthands.h. (See 4d82e11d1b298ecdf969271f2d1da32b658f68fb) I think that runs opposite to these rules.What does everyone think?
NK_INLINE
good or bad?