Re: [Caml-list] newbie: how to call a function with multiple parameters?

[asmadeus77]

Hello,
semicolons are separator inside sentences, that is, just like in C
(you've had theses right)
double semicolons are to tell the compiler you're ending a sentence;
theses can be ommited when there is no ambiguity in the syntax, that
is when there is another "main" let after it. Here, you need them on
line 4.

As for arguments, ocaml works by passing one argument, then gives back
another function, then takes another argument, returns another
function, etc... If you wanted to have a tuple, you'd define the
function with "let print_logic (a, b) ="
Here, since you've defined it with "a b", you need to call it with a
THEN with b. That is : "print_logic true false"

Hope this helps,
Asmadeus

___
Caml-list mailing list. Subscription management:
http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
Archives: http://caml.inria.fr
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

Re: [Caml-list] functions' recursive construction

[asmadeus77]

Hello,

I am not sure I remember what was said exactly, and I'm quite too lazy
to check, but I remember that someone gave a possible implementation
with Obj.magic ?
This also is about the same as the Printf functions which can take a
variable amount of arguments and need a bit of tinkering, and I
remember a function 'eat", taking an integer n, which would ignore the
n first arguments given to the function. Now, if you have this,
getting the first argument should be possible after looking at the
printf source :) (although I could not find it searching in my mails,
sorry)

Actually, I've just glanced at the source and there _is_ a "get_arg"
function, taking a single integer and returning an Obj.magic'ed item
which would most likely be the nth argument of a function - however I
can't separate the function from the rest of the code properly in just
a minute, but it definitly is a place to look at if you need such a
function... Although it might be quite dangerous. But as it was said,
the type is impossible to represent and magic is probably necessary
without typing multiple calls to functions manually; but Jeremy's
solution could be enough :)

Good luck,
Dominique Martinet

___
Caml-list mailing list. Subscription management:
http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
Archives: http://caml.inria.fr
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

Re: [Caml-list] picking / marshaling to strings in ocaml-revision-stable way

[asmadeus77]

Hello,
You can try ocaml sexplib, which use lisp-like structures to store
data... And I don't think this will change anytime soon :)

(I don't know if it can store anything "worse" than a 3-uple, but it
works with pairs and triples, lists, and arrays, which should be
enough for you)

Regards,
Dominique Martinet

___
Caml-list mailing list. Subscription management:
http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
Archives: http://caml.inria.fr
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

Re: [Caml-list] OCaml-ssl and input_string

[asmadeus77]

Hello,
Well, I don't know how it should work, but from what I've seen, it
seems that its read function will wait untill there is something to
read if nothing is available.

Anyway, I have written an input_line (read character by character
untill a '\n') function which allows me to stop as soon as I can match
the begining of the line read with a pattern I get on the greeting.
That will work untill someone tries to send me funny mails with lots
of line feeds followed by weird 8-10 characters strings followed by a
space, I guess it's safe enough.

I just don't feel confident writing things such as :

  let ret = Buffer.create 1024 in
  let str = ref (input_line s) in
  while try Scanf.sscanf !str (Scanf.format_from_string (id^ " ")  "")
(fun _ -> ()) (); false with _-> true do
Buffer.add_string ret !str;
str:=input_line s
  done;
  Buffer.contents ret




On 5/6/08, Berke Durak <[EMAIL PROTECTED]> wrote:
> Hello,
>
> From what I've inferred from ocaml-ssl's documentation, its "read" function
> has Unix-like non-blocking
> semantics.
>
> This means that read returns the number of bytes available, not the number
> of bytes requested.
>
> Hence you need an extra buffering/parsing layer on top of SSL sockets.
> However you are using
> IMAP, which is line-oriented, I guess Ssl.input_string should do what you
> want.
> --
> Berke
>

___
Caml-list mailing list. Subscription management:
http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
Archives: http://caml.inria.fr
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs

[Caml-list] OCaml-ssl and input_string

[asmadeus77]

Hello,

I'm making a simple mail client for gmail using IMAP, and I need to
use ssl sockets.
So far so good, there's a library called ocaml-ssl which works quite
well, untill you need to read what the server sends.

At first, the function input_string in the library is defined by
let input_string ssl =
  let bufsize = 1024 in
  let buf = String.create bufsize in
  let ret = ref "" in
  let r = ref 1 in
while !r <> 0
do
  r := read ssl buf 0 bufsize;
  ret := !ret ^ (String.sub buf 0 !r)
done;
!ret

Which doesn't work, since read will wait untill there is something to
read if there is nothing waiting to be read, thus will never return 0.
I've changed the loop condition to "while !r = bufsize", initializing
r with bufsize, and I've been satisfied with the result during my
early testing (discovering the IMAP protocol in toplevel)

Now, though, I've encountered a problem - it seems that the function
returns too early, meaning that there is still some text left to be
read after it has been invoked.
I've discovered that the size of the returned strings is quite often
1400 or 648, but it can be anything, and changing the variable
"bufsize" doesn't help (I guess it's the size of the packets sent by
the server, but couldn't the socket hold more characters ? There's
usually been a couple of seconds between two input_string calls, so
there should be more than just a couple of KBs)

I've tried adding a poll - (same function with "let fdescr_of_ssl =
file_descr_of_socket ssl in while Unix.select [fdescr_of_ssl] [] []
10. <> [],[],[] do") - but the result, though better, isn't granteed
unless I set the timeout to at least 5-10 seconds, and even there I've
had cases of uncomplete returns, even though they were rare enough to
bear with.

Anyway, I've decided to program this mail client to avoid waiting for
the slow gmail web view, so I'd like to not to use a select there :

Is it possible to read the socket untill a known string, to be sure
the request is done, without waiting after it ? (the string I need to
match is "known_identifier_without_space OK Success\r\n" eventually
leaving the "OK Success" as wildcards to check for (it could be "BAD
why it is bad\r\n"))

I can assume I'm not making two requests at once, thus being sure that
I can read the socket by large chunks untill the end matches, but I
would rather not do so if possible.

Any clue ?

Thanks,
Dominique Martinet

PS : A friend of mine is wondering if there's a regexp for ranges in
the Str module (like "?" would be "a{3-4}" in pcre), I'm sure
he'll be glad if someone could ensure him there isn't any :P

___
Caml-list mailing list. Subscription management:
http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
Archives: http://caml.inria.fr
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs