Jonas Lund

> * I am split on whether I am willing to support a proprietary dialect in the first place. Well there's PicoLove that's an opensource impl for porting and I'm...

As to the errornous number suffix i think there was some version information in some of Apples iOS headers that was in a form similiar to 5_1_0 and only used...

Again, not remembering everything exactly but i think that it was without the "case NL" around line 743 in commit(d67cee2) that macro arguments spanning over multiple lines broke since the...

I think there was something that made working with the VFS hard in this case and i think i reasoned that the abstraction would be to allow compile servers that...

(As the INVALID vs IDENTIFIER thing, i think i initially changed to IDENTIFIER when i didn't notice what was wrong but changed it back to invalid in a4ce2c8 )

The code for the frameworks seemingly looks fine to me, simpler than my variant also that was overly defensive and kinda went in the wrong way (I think my variant...

hashing on virtualfile _sounds_ more sane but i remember some weirdness described somewhere (as to why i separated the import variants, i'll try to find the source for this).

Actually, i think that you're right about just going with the VFS identity. I separated and "" because i didn't think i could rely on simple names (and didn't know...

I was pondering about this (and with other tools also) since i often find myself iterating things in different projects with different needs usually resorting to a makefile or batch...

Ok investigated it a bit more, made a minimal test case and found out that it seems to be a compiler bug inside MSVC 2008 and forward (2005 compiles fine)....