On Fri, Nov 30, 2001 at 04:16:40PM +0100, Andre Poenitz wrote: > On Fri, Nov 30, 2001 at 03:12:02PM +0000, Jules Bean wrote: > > There is a standardised 64 bit integer type: long long int is standard. > > It's in C99 as far as I know, but in C++98?
I don't know. It's not in Stroustrup; but I can't see any such details about data size there. *sigh* Portability is not obtained by making 'int' represent 'a convenient size': rather, portability is obtained by knowing exactly how many bits you have to play with on any system... > > Anyhow, this is all (fortunately) moot: you'll notice that that wasn't > > consistent with the rest of my message, where I say that there are > > 65536 sp in a point: i.e. sp is 2^(-16), and I made a mistake ;-) > > > > TeX uses a 32 bit int, measured in sp, for all its calculations AFAIK. > > So having "int" would be fine? That is my belief. Although, when you care about these things, I've always thought a naming scheme like u32int is better style. Jules
