caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Basile STARYNKEVITCH <basile@starynkevitch.net>
To: Oliver Bandel <oliver@first.in-berlin.de>, caml-list@inria.fr
Subject: Re: [oliver: Re: [Caml-list] Unix.chmod => what codes for what mode?!]
Date: Fri, 23 Apr 2004 23:08:52 +0200	[thread overview]
Message-ID: <20040423210852.GA1294@ours.starynkevitch.net> (raw)
In-Reply-To: <20040423200949.GD271@first.in-berlin.de>

On Fri, Apr 23, 2004 at 10:09:49PM +0200, Oliver Bandel wrote:
> 
> To: Richard Jones <rich@annexia.org>
> Subject: Re: [Caml-list] Unix.chmod => what codes for what mode?!
> 
> On Fri, Apr 23, 2004 at 07:01:02PM +0100, Richard Jones wrote:
> > On Fri, Apr 23, 2004 at 07:16:34PM +0200, Oliver Bandel wrote:
> > > Hello,
> > > 
> > > The Unix.chmod function is bad documented.

No it is not

> > 
> > Almost certainly you're not using octal numbers right.  The format
> > for octal numbers in OCaml is different from C, for example:
> > 
> > Unix.chmod filename 0o755
> > 
> > would have the same effect as:
> > 
> > % chmod 755 filename
> 
> 
> Would be nice if this litte more of explanation could be added to the
> reference Manual. 

I respectfully mostly disagree. Unix.chmod Ocaml function is, like the
chmod C function, of type
   int -> string -> unit

The way a language represent integers is not related to Unix.chmod;
consider for example an imaginary programming language where you have
to express number in Roman notation. Then, you should have to code
  chmod (CDXCIII, "/your/path")
or perhaps
  chmod (CCCCLXXXXIII, "/your/path")
or maybe
  chmod (IIVD, "/your/path")
however I am not sure that IIVD is a correct latin notation for
four-hundred-ninety-three which is written 0o755 (as an octal number)
in Ocaml, ahd which does the same as the shell command 
    chmod u=rwx,g=rx,o=rx /your/path 

My point is that every experimented Unix programmer knows that the
integer argument to chmod is actually a bit mask. It is not the
purpose of the Ocaml documentation to remind this (otherwise, the
Ocaml documentation for the Unix module should be as big as the
section 2 of Unix man pages, which would triple its volume).

In other words, when I code in Ocaml a Unix call and I forgot what it
does, I consult the man page for the system call to understand what
the syscall does, and then I read the description of Unix module in
the Ocaml documentation.

However, per your suggestion, I added in unix.mli (at work, on the
CVS of Ocaml) the above comment for type file_perm = int definition 

(** The type of file access rights, e.g. [0o640] is read and write for
user, read for group, none for others *)


Perhaps the only functions from the Unix module which could be a bit
more documented are the basic Unix.read and Unix.write functions,
because Unix.read truncate the amount of bytes (see
otherlibs/unix/read.c) read to UNIX_BUFFER_SIZE which is 16384 (in
unixsupport.h), while Unix.write actually loops and may issue several
write system calls, by chunking the written string in (possibly
several) chunks of UNIX_BUFFER_SIZE bytes. So (as Didier Rémy told me)
if a Unix.write of a million byte string is interrupted, you might get
a Unix_error exception after a hundred thousand bytes have been
successfully written. However, I don't know exactly what is the
recommended Posix behavior (w.r.t interrupts) for a read system call
(in C) of a million byte.

The reason of this chunking is that Unix.read and Unix.write actually
copy the data into (or from) an internal buffer before making the
(interruptible and preemptible) system call. It is important to have
the syscall be blockable, so that another Caml thread could run (hence
allocate and call the GC, this is why the data is copied).


In practice, it might be better to use Unix.read & Unix.write with a
length of no more than 16384.

-- 
Basile STARYNKEVITCH         http://starynkevitch.net/Basile/ 
email: basile<at>starynkevitch<dot>net 
aliases: basile<at>tunes<dot>org = bstarynk<at>nerim<dot>net
8, rue de la Faïencerie, 92340 Bourg La Reine, France

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners


  reply	other threads:[~2004-04-23 21:08 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-04-23 20:09 Oliver Bandel
2004-04-23 21:08 ` Basile STARYNKEVITCH [this message]
  -- strict thread matches above, loose matches on Subject: below --
2004-04-23 20:09 Oliver Bandel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20040423210852.GA1294@ours.starynkevitch.net \
    --to=basile@starynkevitch.net \
    --cc=caml-list@inria.fr \
    --cc=oliver@first.in-berlin.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).