Reply to thread

Message: [QUOTE="Richard Heathfield, post: 1691939"] Oh, I see. Since you posted it in comp.lang.c, I made a not unnatural, but nevertheless apparently incorrect, assumption that the code was intended to be portable. If it were not so intended, then of course it's not surprising that I got compilation errors. What /is/ surprising is that you should post (a link to) such code here in comp.lang.c. Yes, I managed to get a sieve working cheerfully even right up to one million even under MS-DOS (admittedly I saved lots of space by testing 2 separately and throwing out all other even numbers). Why not just use bignums, which can be written 100% portably and with 100% accuracy? Solutions are possible that are correct and portable even to such platforms. [/QUOTE]