16:02 <er> match (x) with | x1->stmt1; | x2->stmt2; stmt3;

16:03 <er> does stmt3 associate with the match with x2, or all matches?

16:03 <er> how does one force it to associate with the parent?

16:05 <Saulzar> I believe it is the same thing as match (x) with | x1->stmt1; | x2-> (stmt2; stmt3;)

16:06 <Saulzar> But using x; is also equivalent to let () = x in .. I think, if that helps

16:07 <er> Saulzar: that's what the output of my program suggests too.

16:07 <er> Saulzar: but I'm trying to make it associate with all the matches, so...

16:07 <Smerdyakov> Saulzar, not quite equivalent at the level of concrete syntax. Different associativity.

16:07 <Smerdyakov> er, use parentheses or 'begin..end'.

16:07 <er> (match x with | x1->stmt1;|x2->stmt2) stmt3;

16:08 <Smerdyakov> Almost. That's a function application.

16:08 <Saulzar> Ah ok - I'm still not 100% on it myself ..

16:09 <er> Smerdyakov: um, exactly. that's what ocamlc says as well, so I'm glad you agree...

16:09 <er> :)

16:11 <er> what is the correct way?

16:12 <Saulzar> Semi-colon the brackets

16:12 <Smerdyakov> er, add a semicolon after the close-paren.

16:12 <Smerdyakov> er, there is no magic or complicated rule to learn here. OCaml has only one class of program elements at this level: expressions.

16:13 <Smerdyakov> er, calling something a "statement" will get you smacked. ;)

16:13 <Saulzar> Yeah, it just seems there's something wierd going on when you first see there is also syntax for imperitive stuff

16:13 <Smerdyakov> er, so you operate on 'match' expressions with operators just like you would with integer constants.

16:13 <Saulzar> Hmm, I think I prefer begin/end to brackets... makes it clear you're using imperitive style

16:13 <Smerdyakov> Saulzar, of course, there is no objective reason to say it makes that clear, since they're just syntactic "anti-abbreviations" for parens.

16:14 <Saulzar> Yeah, I know..

16:15 <er> i called them statements because they cause side effects

16:15 <Smerdyakov> er, I don't think that is standard usage.

16:16 <er> ok

17:07 <moea> is there a convenient way to convert floats into their internal ieee 754 representation?

17:07 <moea> something like python's struct.pack('!d', float)

20:45 <mikeX> what's the deal with GWindow.message_dialog in lablgtk2? I can't understand that [> DELETE_EVENT] as a') buttons at all :/

21:26 <mikeX> ok let me put it this way, what could "([> `VARIANT1 ] as 'a) `VARIANT2" possibly mean?

21:27 <mauke> 'a looks like a type variable

21:28 <mikeX> hmmm :/

21:29 <mauke> see http://caml.inria.fr/pub/docs/manual-ocaml/manual006.html

21:29 <mauke> "Polymorphic variants"

21:29 <mikeX> yeah I am, not that I can make that much of it :(

21:30 <mikeX> >looking at it<

21:36 <mikeX> can you think of something that matches this type?

21:37 <mauke> is that the whole type?

21:39 <mikeX> well the actual type is "([> `DELETE_EVENT ] as 'a) buttons", with "type buttons = [ `CANCEL | `CLOSE | `NONE | `OK | `OK_CANCEL | `YES_NO ]"

21:39 <mauke> huh? why "as 'a" then?

21:40 <mikeX> :?

21:40 <mauke> if 'a isn't used anywhere else, why is it declared there?

21:40 <mikeX> don't ask me :/

21:46 <mikeX> *sigh*

21:46 <mauke> is this code online somewhere?

21:47 <mikeX> yeah sure, hold on

21:47 <mikeX> http://oandrieu.nerim.net/ocaml/lablgtk/doc/Gtk.Tags.html#TYPEmessage_type

21:47 <mikeX> here's the "buttons" declaration

21:48 <mikeX> hmm, I can't seem to find the other one online

21:49 <mikeX> http://oandrieu.nerim.net/ocaml/lablgtk/doc/GWindow.html

21:49 <mikeX> look at 'GtkMessageDialog'

21:50 <mauke> that's not the same "buttons"

21:52 <mikeX> oops

21:52 <mikeX> :/

21:52 <mauke> it seems to declare a parameter labeled "buttons" of type ([> `DELETE_EVENT ] as 'a) buttons

21:52 <mauke> 'a is declared there because it's used in the return type

21:53 <mikeX> so, what is the type of the parameter (an example)?

21:53 <mauke> hmm, "buttons" seems to be an abstract type

21:55 <mauke> let's make a dummy definition: type 'a buttons = Whatever of 'a;;

21:55 <Smerdyakov> Polymorphic variants are an atrocity.

21:55 <mauke> then Whatever `DELETE_EVENT;; results in - : [> `DELETE_EVENT] buttons = Whatever `DELETE_EVENT

21:57 <mauke> I've never used them

21:57 <mikeX> :/

21:58 <mikeX> i see (i think)

21:58 <mikeX> thanks for the help

22:51 <Amorphous> Saulzar: some days ago you wrote "I've noticed something about using mutable values... eg. as soon as I use a Hashtbl in a data structure then everything from there on down is 'infected' with being mutable" i'm interested in an example for that, if that wouldn't bother you, could point me to one, please?