4tH compiler

I'll see if I can fix this. Shouldn't be too difficult. Thanks for testing it! Hans Bezemer

Well, since the introduction of [IGNORE] the overhead has gone; no penalty at all. Amazing how fast I came to it - I just analysed the options one by one by clearly stating the design objectives and it came to me almost automatically.. Anyway, I was leafing through the manual when I bumped into FSL-UTIL.4tH

Hi 4tH-ers! Ok, this got to be thoroughly tested, but I found a way to make 4tH IGNORE symbols. It works like this: the [IGNORE] directive adds a symbol containing the token STRINGD (usually issued in conjunction with ,") with a NEGATIVE argument (which isn't possible). The peephole optimizer now acts

Op zaterdag 17 maart 2012 11:20:15 UTC+1 schreef The Beez het volgende: I also added a dummy ZenFP definition to "fpow10.4th" for ME>F. It will create a little overhead, but may be a life saver if you want to create programs that run under both ZenFP and ANS FP. Hans Bezemer

Ok, I found it. It's called "binary scientific notation". It looks so bizarre, because it is "normalized", that is: scaled so the mantissa is as big as possible (means: topmost digit is set). Converting a decimal mantissa/exponent to this format isn't entirely trivial (see: [link]).

Don't expect an "ME>F" anytime soon. As I expected the ANS Float FP format is rather bizarre. Here is 60e0 for you (mantissa (bin), exponent (bin), float): 111100000000000000000000000000 000000000000000000000000000000 0000 111111111111111111111111100011 0 60.0000000000000000 The exponent seems to be in format mantissa*2^n rather than

OK, I'm done. All "S>FLOAT" constructs have been removed from the libraries, unless they're used to initialize tables. They're called only once, no need to fix that. Yes, readability suffered in some areas, but even if I come up with a better ... [More] solution I don't it will change that much: - Even a ME>F word (Mantissa/Exponent to Float) word won't help where [Less]

Many thanks for your thoughtful reply, Hans. The Latex words look quite interesting. I'll dig deeper. Thanks, LRP -----Original Message----- Sent: Thursday, March 15, 2012 6:02pm To: [email protected] First of all, I am only the ... [More] maintainer of 4tH and in no sense any better than any other representative of the Forth community, but I will try to answer [Less]

First of all, I am only the maintainer of 4tH and in no sense any better than any other representative of the Forth community, but I will try to answer this as unbiased as possible. First, pure C (C99) has AFAIK no network facilities as well. The only framework for C is derived from the POSIX standard. ANS Forth has neither,

Hi Hans, I've been following 4th out of the corner of my eye for some time. But your recent posts have pulled 4th into more direct focus. I just re-read the intro to the 4th manual and must say, "Outstanding!" Thank you for all this fine work. A couple of questions--- but first, some background to put the questions in context.