06:36 <flux__> I wonder if type inference would break if ocaml supported implicit downcasting..

07:50 <andreas__> hi

07:55 <andreas__> Does OCaml have global variables? I would like to have a global configuration record which is accessable from within all my modules.

07:58 <flux__> sure (although you might want to consider it thoroughly before doing it ;)) : let config = ref None

07:58 <flux__> and then Modulename.config is your configuration

08:00 <andreas__> You mean putting the configuration in its own Module?

08:00 <flux__> yes

08:01 <flux__> there are no globals that are automatically opened everywhere, except for Pervasives

08:01 <flux__> s/globals/modules/

08:01 <flux__> if that's what you were looking for

08:01 <andreas__> what do you mean by s/globals/modules ?

08:02 <flux__> I meant modules when I wrote globals

08:02 <andreas__> oh, ok

08:02 <flux__> it is a sed-expression, often seen in (informatin) technically oriented channels

08:03 <andreas__> you mean regular expression ;)

08:03 <flux__> well, a sed-expression may contain a regular expression

08:03 <flux__> and they s-command indeed does, as its first argument

08:03 <flux__> but it's not a regular expression in whole

08:04 <andreas__> thx, for the help concerning globals - I will try it

16:42 <mikeX> what's the overhead of wrapping the 'main' function of a program in a 'try .. with' statement? is it significant?

16:46 <Schmurtz> I think it's unsignificant

16:47 <mikeX> i see, thanks

16:47 <Schmurtz> in general, "try“ statements only take time when entering them, and when raising an exceptino

19:35 <mikeX> This is related to exception checkers that exist at least for OCaml. The external tool for OCaml is both transparent (i.e. it does not require any syntactic annotations) and facultative (i.e. it is possible to compile and run a program without having checked the exceptions, although this is not suggested for production code).

19:35 <mikeX> i've never seen anything like that, is that true?

20:02 <love-pingoo> pango: the other day you told us about a nice statically typed scripting language

20:02 <love-pingoo> if you still remember the name, I might like to take a look

20:04 <pango> http://compsoc.dur.ac.uk/kaya/ probably

20:06 <love-pingoo> thx

20:22 <love-pingoo> compiled scripting language.. hmm, I wonder how one may define "scripting"

20:22 <zmdkrbou> hehe

20:23 <zmdkrbou> strange idea

20:23 <lde> there is also no such think as `compiled language'

20:23 <love-pingoo> another thought is that kaya is another user of the boehm GC.. if I'm not mistaken, boehm-gc is a hack :(

20:23 <lde> *thing

20:24 <pango> hack as in "conservative gc" ?

20:24 <love-pingoo> hack as in "guess what's a pointer"

20:24 <love-pingoo> if I'm not mis-informed

20:24 <pango> yes, that's it

20:25 <pango> maybe it's not the correct english term

20:25 <love-pingoo> I guess boehm is quite efficient however, to be so popular

20:26 <pango> or because it's conservative ;) you can use it in C with no modification

20:28 <love-pingoo> yeah.. holy lazinezz

20:29 <love-pingoo> anybody knows whether the Genlex module's lexers are reentrant ?

20:33 <laurence> love-pingoo: Also, I'm told, because precise collectors don't agree well with GCC's internals. At least that's the reason for Boehm being used in GCJ.

20:34 <mellum> well, there is no reason for the code generated by gcc to have anything to do with the internals of gcc

20:35 <mellum> boehm is probably used because it's easiest. You need to do basically nothing.

20:36 <laurence> mellum: Yes, there is (got this from the lead for GCJX). Using a precise collector with GCC apparently would mean a heavy re-write of the back end.

20:37 <laurence> I'm not talking about GCC's operation. I'm just talking about generated code.