develooper Front page | perl.perl5.porters | Postings from November 2000

Re: PerlIO - what next ?

Thread Previous | Thread Next
From:
Nicholas Clark
Date:
November 8, 2000 06:24
Subject:
Re: PerlIO - what next ?
Message ID:
20001108142416.A98736@plum.flirble.org
On Wed, Nov 08, 2000 at 01:50:06PM +0000, Nick Ing-Simmons wrote:
> This means knowing what it "should" look like to perl5:
> 
> Do you want getc($fh) return character or a byte depending on the handle's 
> top "layer" ?
> 
> Likewise does 
> 
>    $count = read($fh,$buffer,10); 
> 
> read 10 chars or 10 bytes depending on mode ? 

What do you mean by mode? "sum of effects of layers on the handle"
{
  use utf8;	# And if I understand it correctly this will become a no-op
  Surely there's no difference between a byte and a character?
}

> How do you want I open a file which is utf8 or shift-jis or whatever?
> 
> What should "binmode" mean now we have layers?

I'd expect [my opinion] that getc($fh) gets scalar of length 1, and the read
call returns a scalar of length 10.

A given (disk) file is treated as utf (and faulted if bad utf?) and stuffed
into a scalar with characters > 255 if there's
  (a layer on the handle to convert the raw bytestream to utf) ||
  (the handle is using an OS API that returns utf)

If there's no layer doing conversion then the raw disk bytes end up in
a perl scalar which only contains characters < 256 independent of
whether that byte sequence is a valid utf encoding (of some other
sequence)

I'd agree with Simon that what layers get stacked is based on use open;

Nicholas Clark

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About