Am Dienstag, 19. September 2006 23:05 schrieb Christian Ohm: > Hm, I wouldn't have changed 'ip', as that name is used in other places > as well, but OK. Maybe change those other places, too then? ;) I had quite some difficulties imagining what "UWDORD*ip" could be, till Troman told me that it is an Instruction Pointer. So I felt like changing that.
> The attached patch fixes g++ compilation broken by Troman's commit. The > removed casts inspired a new SVN rule: "ALWAYS do 'svn up' before 'svn > diff' (and fix conflicts, if necessary)." I guess I should make a new > draft for the SVN guidelines soon. I guess everybody is doing that. Just MSVC doesn't check the code as good as GCC does, and G++ does it even better than GCC. > The removed inline was necessary for g++ to actually find the function > when linking. "extern inline" doesn't really make sense. Oh, that does make sense. I even got it from a website, because I wondered if it is possible. http://www.greenend.org.uk/rjk/2003/03/inline.html But that could be a GNUC dialekt... > If we want to > keep the function inlined (probably not necessary, as the compiler will > inline functions when optimizing*), we can add it to > lib/script/script.h, but then we have to include the RetStackPos > variable there as well. I added those inlines when I replaced if(retStackPos < 0) with if(retStackIsEmpty()). I also had a look at the other retStack functions and choose several to be inline. Then I detected that most of them are only used in interp.c and thuse moved them over and made them static inline. retStackCallDepth was the only remainder. --Dennis
pgpAmqWwM1KgE.pgp
Description: PGP signature
_______________________________________________ Warzone-dev mailing list [email protected] https://mail.gna.org/listinfo/warzone-dev
