Catherine
Catherine
I agree that the interface you propose addresses the design constraints you list pretty much perfectly, but I think it's far too heavy to be usable. Personally I can't stand...
Yes. But also, types are not their own end. Types should help the programmer by exposing bugs, not hinder them by requiring to work around interfaces hostile to common cases....
Hmm, that looks vaguely promising, but I don't really understand how one's supposed to use it. Can you show some example?
I mean--`result` is still abstract, this is the same basic problem, just shuffled around.
That doesn't help at all. E.g. if we have a `get_width : string → int fn` and it's used in an expression `let is_wide = get_width "window" > 1000` then...
Oh, I see your overall approach now, now that I look at `val (>>=) : 'a fn -> ('a -> 'b fn) -> 'b fn`. I concur that this is...
This seems wrong. What exactly is the issue here?
Known issue, not considered important enough to spend time on. Patches welcome, of course.
In theory, it should be a host-runnable binary, i.e. the same as `ocamlfind doc`. I don't recall what was wrong to make it a Windows one, some build system plumbing...
It immediately crashed when I tried to build it on `4.02.3+32bit`. Further the camlimages' OMakefile does not do everything using ocamlfind, and generally tries to do annoying things like unconditionally...