00:10 <Peaker> Smerdyakov: the operator. syntax means its a float-op?

00:10 <Smerdyakov> That's the convention for the built-in floating point operators.

00:10 <Smerdyakov> It's not codified in the language in any other way.

00:11 <Peaker> I see in WP's example that they use Printf.printf, but there's no "import" of Printf. Is it a module? how is it imported? is the case significant?

00:11 <Smerdyakov> Printf is a module, and case is significant.

00:12 <Smerdyakov> In the more traditional compilation mode, modules are made available with command-line flags.

00:12 <Smerdyakov> (Like libraries in C)

00:12 <Peaker> Ah. Just recently I realized that I don't like Python's import statement, so this may be cool :)

00:12 <Peaker> Camel case means..?

00:13 <Smerdyakov> likeThisHere

00:13 <Peaker> No no, I mean what is its semantic meaning in ocaml?

00:14 <Peaker> I meant above: "is ocaml case significant", not "is ocaml case sensitive"

00:14 <Smerdyakov> None. The only significant of case is that the first character of an identifier is uppercase for modules, signatures, and constructors, and the first character is non-uppercase for the rest.

00:14 <Peaker> ah, ok

00:15 <Peaker> # does not denote comment?

00:16 <Smerdyakov> Nope

00:16 <Peaker> but #! is ignored?

00:16 <Smerdyakov> If it is, I wasn't aware of that. Maybe that's true in the first line, to support UNIXisms.

01:07 <ayrnieu> I'd like to write a networking application for a windows-using friend, which application needing a very simple graphical interface (two listboxes and a button with complaining dialogues and perhaps scrolling diagnostic messages would do). I'd also like to be able to pass on a small zipfile containing all the program requirements.

01:07 <ayrnieu> I only have Solaris machines for development. How appropriate for this task is O'Caml?

01:08 <Smerdyakov> I think OCaml works fine on Solaris, and labltk seems to work fine in Cygwin.

01:09 <ayrnieu> I'm sure O'Caml works fine on Solaris; I'm worried about how I can provide binaries that run on Windows.

01:09 <Smerdyakov> I've never done cross-compiling.

01:09 <ayrnieu> I see. Well, I'll wait to get a windows machine.

02:19 <Revision17> O'Caml works 100% fine on solaris

02:19 <Revision17> solaris 9 anyhow

02:19 <Revision17> I've only used stdlib and extlib with it though

03:05 <KrispyKringle> Hey folks.

05:49 <TylerE2> Is the performannce of the complex module decent? Or am I better of using e.g. a float * float and writing my own ops? All I'll be doing is adds and mults

06:19 <TylerE2> anyone awake?

06:23 <m3ga> TylerE2: a little

06:23 <TylerE2> ;)

06:30 <pango__> the Complex module looks 100% ocaml, see your lib directory for the sources...

06:38 <pango> with some inlining (ocamlopt -inline 10 ?) performance should be decent... In the end, doing some benchmarks is the only way to be sure...

06:58 <TylerE2> ugh

06:58 <TylerE2> I'm getting bit by mutable arrays

06:59 <TylerE2> ?let points = Array.make size (Array.copy (Array.make size 0));;

06:59 <TylerE2> which should work, but the arrays are still sharing

07:04 <TylerE2> n/m figured it out

07:04 <TylerE2> make_matrix

07:04 <pango> use Array.init (or directly Array.make_matrix)

07:05 <pango> in your code Array.copy is only done once (and the inner Array.make array will be eventually discarded... you gained nothing)

07:05 <TylerE2> doh

07:05 <TylerE2> I'm too used to lisp :P

07:05 <pango> Array.init takes a function, that's evaluated once for each element... only way (afaik) to prevent aliasing

07:05 <TylerE2> ahh, ok

07:06 <TylerE2> the standard library doc is a bit...dense

07:06 <TylerE2> be nice if someone through up a wiki-fied version somewhere

07:06 <pango> of course make_matrix is the way of the lazy ;)

07:09 <pango> let points = Array.init size (fun _ -> Array.make size 0)

14:39 <Smerdyakov> exa, those messages of yours are annoying. :P

14:40 <exa> It's a feature of Konversation :)

14:40 <Smerdyakov> I wish you would disable it for this channel.

14:51 <exa> Yea

14:56 <KrispyKringle> Konversation. Geeze. You just know the KDE guys are Anarkists, too.

14:56 <KrispyKringle> So anyway, I can use references as mutable variables in a sort of global scope, right?

14:57 <KrispyKringle> Like, I can do "let stack_tops : 'a list ref = ref []" and then later, stack_tops := somenewvalue, yes?

14:57 <smimou> yes, but

14:58 <smimou> it's going to be a '_a list

14:58 <smimou> which means that this list can only contain values of one type

14:58 <smimou> (and btw you don't have to specify explicitely the type)

14:59 <KrispyKringle> Wait, aren't lists usually homogenous types?

14:59 <KrispyKringle> I'm not sure what you mean.

14:59 <smimou> sorry I wasn't clear

14:59 <smimou> list are homogenous

15:00 <smimou> but you can't do

15:00 <smimou> stack_tops := [1];;

15:00 <smimou> and later stack_tops := ["toto"];;

15:00 <KrispyKringle> Right.

15:00 <KrispyKringle> I thought that's what you meant.

15:00 <KrispyKringle> OK, so I have:

15:00 <KrispyKringle> ooh, wait a sec

15:01 <KrispyKringle> Yeah, so I have let stack_tops = ref [];; let stack_tops_next = ref [];;

15:01 <KrispyKringle> and in one o fmy functions, I do stack_tops := next_stack_tops;

15:01 <KrispyKringle> next_stack_tops := []

15:01 <KrispyKringle> And that gives me a type error. This expression has type 'a list ref but is here used with type 'b list

15:02 <KrispyKringle> But given that the two lists aren't yet used anywhere else, it should infer that they're the same type, no?

15:02 <KrispyKringle> I tried putting it explicitly, but no use.

15:02 <Smerdyakov> Do you know the typing rule for :=?

15:02 <smimou> you forgot a ! I guess

15:03 <KrispyKringle> Smerdyakov: This expression has type 'a list ref but is here used with type 'b list

15:03 <KrispyKringle> er

15:03 <KrispyKringle> wrong clipboard

15:03 <smimou> (stack_tops := !next_stack_tops;)

15:03 <KrispyKringle> d'oh

15:03 <KrispyKringle> yeah.

15:03 <KrispyKringle> thanks.

15:03 <Smerdyakov> smimou, I think it would have been more helpful to let _him_ figure that out.

15:03 <KrispyKringle> thanks. I just missed that.

15:03 <KrispyKringle> Smerdyakov: I knew the type for :=

15:03 <smimou> sorry

15:04 <smimou> let's forget what I said :)

15:04 <KrispyKringle> It's just early in the morning. ;)

15:05 <Smerdyakov> KrispyKringle, if you know the types for all aspects of your program, then you can always systematically determine the meanings of error messages.

15:09 <KrispyKringle> Smerdyakov: Sure. But I'm human. I make mistakes.

23:20 <exa> i feel that more power is being drawn to ocaml

23:32 <dylan> hmm>

23:32 <dylan> *?

23:50 <Raziel> hi

23:51 <Raziel> I have a problem while trying to compile a program that uses Num library

23:56 <Smerdyakov> Read the manual.

23:57 <Smerdyakov> And it's properly called "the num library." (lowercase)

23:58 <Raziel> hum.

23:58 <Raziel> it's weird, two days ago it was impossible to compile the program

23:59 <Raziel> but now it compiles without problem

23:59 <Raziel> :\