01:34 <taw> are there sobe deconstructing when clauses ?

01:34 <taw> | Foo x when y = Bar z -> ???

01:39 <mrvn> ???

01:40 <mrvn> Whats y supposed to be and whats your question?

01:40 <mrvn> and whats z?

01:41 <mrvn> There should be at least another x in there

01:41 <taw> y some variable

01:41 <taw> z - binding for contents of y

01:41 <taw> of course my syntax is just example

01:42 <mrvn> z doesn't get bound there.

01:42 <taw> | Foo x when (match y with Bar z -> true | _ -> false) -> ???

01:42 <taw> how to get z bound on the right side ?

01:42 <mrvn> s/z/_/

01:43 <taw> that's probably a better question

01:43 <mrvn> match (a,y) with (Foo x, Bar z) ->

01:43 <taw> repeating that sure works

01:43 <taw> mmm

01:43 <mrvn> If you need the z later

01:44 <taw> putting another var in match for each such when kinda sucks

01:44 <taw> ah, whatever

02:46 <mrvn> # let print_s = Printf.printf "s='%s'\n";;

02:46 <mrvn> s='val print_s : string -> unit = <fun>

02:46 <mrvn> # print_s "1";;

02:46 <mrvn> 1'

02:46 <mrvn> - : unit = ()

02:47 <mrvn> Is that intentional?

02:49 <mrvn> I would expect print_s to be a function that prints s='%s' each time.

02:52 <baader> hm

03:06 <Verbed> mrvn: That's just not how printf works.

03:06 <Verbed> I'm not familiar with O'Caml so rendering an explanation is difficulty..

03:08 <Verbed> The function doesn't curry as normal. It scans through the format string, printing whatever is immediately accessible, and when encountering a % flag, it returns a function. This function is called with the appropriate argument, which is then printed. This continues until the format string has been processed.

03:09 <Verbed> It acts so strangely due to 'magic' calls to the OCaml type system.

03:09 <Verbed> Otherwise such a function would not be typable.

03:10 <Verbed> ( typeable )

03:12 <Verbed> It would be appropriate to call it a kludge.

03:12 <Verbed> I believe the new beta adds true support for format types.

03:16 <Smerdyakov> Template Haskell does it even better. :-)

03:16 <Verbed> I favour Scheme anyways.

03:17 <mrvn> Verbed: printf prints the strings after the first % correctly. It could do the same for before the first %.

03:17 <Verbed> Yes, it still won't ask as you want it to.

03:18 <Verbed> s/ask/act

03:18 <mrvn> it would.

03:18 <Verbed> And what exactly do you want from it?

03:18 <mrvn> It would have to do the same it does for "string after %?" for "string before %?"

03:18 <Verbed> I don't understand that statement.

03:19 <Verbed> Perhaps you could illustrate it with some code?

03:19 <mrvn> let print_s = fun str -> Printf.printf "s='%s'\n" str;;

03:19 <Verbed> Let me provide you with some code.

03:19 <mrvn> Thats what it should do.

03:20 <Verbed> Yes, you might do that explicitly if you like.

03:20 <mrvn> Implementation wise I would like an uncurried printf.

03:20 <Smerdyakov> mrvn, the OCaml docs tell you to eta-expand your curried call to get the result you want.

03:21 <Smerdyakov> mrvn, so you should have known. ;)

03:21 <mrvn> yeah, still not what I had expected.

03:22 <Verbed> Do you feel that you understand its behaviour now?

03:22 <mrvn> And converting printf into a printf that acts like an uncurried one but still is curried is pretty complex.

03:22 <mrvn> Verbed: I know why printf does what it does. I just don#t like it :)

03:25 <mrvn> In a setuid prog how would you execute some function of the proram as the actual user (most likely not root)?

03:25 <mrvn> Do I realy have to fork, setuid and waitpid?

03:26 <Smerdyakov> Feh on Unix :P

03:27 <mrvn> Why doesn't ocaml have seteuid bindings?

03:28 <Verbed> It doesn't?

03:28 <Smerdyakov> Perhaps that function isn't standard enough?

03:29 <Smerdyakov> The 2002 SML Basis doesn't have it, either.

03:29 <mrvn> seteuid is conforming to BSD 4.3

03:30 <Smerdyakov> What's the first ANSI or POSIX standard it appears in?

03:30 <mrvn> aparently none

03:31 <Smerdyakov> Question answered ;)

03:31 <mrvn> But I guess any unix system has it

03:31 <mrvn> or at least most

03:32 <Kinners> the qnx docs say POSIX 1003.1a Draft 15

03:56 <mrvn> Does Unix.truncate create the file if non existent?

04:07 <phubuh> mrvn: truncate(3) doesn't

07:42 <two-face> yo