http://math.nist.gov/tnt/
Thanks to Corned Bee who got me mad enough to start looking for iso/ansi C++ specs on the net - free of course.
Printable View
http://math.nist.gov/tnt/
Thanks to Corned Bee who got me mad enough to start looking for iso/ansi C++ specs on the net - free of course.
:);):D
LOL, not what i was expecting :p
yuck #ifdef macros everywhere :p I'd have done it with templates :)
yeah, but thats because you are the template god. :p
well I am :cool: ;)
Kedaman - you don't get it.
#ifdef reduces code size. It's a great way to optimize code.
If something isn't compiled, it isn't in there taking up space for no reason.
On the other hand it is not elegant. Elegance and fast, efficient code are not synonomous. Look at CornedBee's Fibonacci recursion code. Elegant and crashes the system on calculating the 48th fib number, to take an extreme example.
You should see what Kedaman's working on at the moment. It's a neat method of getting both at the same time.
You have your types wrapped up into classes, so that you can't mix constants, for example. But it all gets inlined by the compiler in the same way :)
Pity MSVC can't go all the way on it though :(
Bork it! I founded a new paradigm just last weekend :D i'm still search around the net to see if someone hasn't done something similar yet, but it's kinda hard when you don't know what to look for
keda: they probably want it to work with MSVC++ ;)
Actually, most of what ked wants seems to be working on VC7b1 - so it might be better on the new beta 3.Quote:
Originally posted by CornedBee
keda: they probably want it to work with MSVC++ ;)
Well, actually you haven't seen anything of my newest work parksie ;) you'll be surprised what all you can do, but what i'm really annoyed with now is the lack of partial specialisations of the templates
That should be solved by GCC3...see your other thread.
Despite it needing to be run under Cygwin, I think it produces native executables (i.e. the programs you write don't need it).