Mirar @ Pike developers forum wrote:
Oh right, I spotted that, yes. snprintf doesn't exist there. But could we have some sort of bugfix?
I'd be happy with a %.10g instead of %f. (Anyone against?)
IMO, as a minimum patch I'd suggest:
sprintf(buf, "%.*g", MAX_FLOAT_SPRINTF_LEN, (double)sp[-1].u.float_number);
It's reasonably accurate, and reasonably safe.