On Mon, Jun 10, 2002 at 08:16:12PM +0100, Nick Ing-Simmons wrote:
> Jarkko Hietaniemi <[EMAIL PROTECTED]> writes:

> >Ummm, well, I guess that would work too, but wouldn't the "binmode(FH)
> >means an implicit :bytes (if available)" what you suggested earlier be
> >even better?

I like this best. To me, unadorned binmode means as raw as possible.
(For example; I'd thought that the Storable tests were portable to DOSish
systems, so it was somewhat surprising seeing Jarkko needing to patch them
to make them work properly on Unix)

> Perhaps - I intend to try it and see.
> What bothers me is binmode() after :encoding is bad news if binmode()
> turns off UTF-8 ness:
> 
> open(FH,">:encoding(Xxxx)",$file)
> binmode($file);
> 
> The translate tables expect UTF-8 encoded chars, so blindly turning 
> off PERLIO_F_UTF8 is a bad idea.
> 
> I have a feeling that some of tests do that - to avoid CRLF issues.

Can we write that 

open(FH,">:encoding(Xxxx):lf",$file)

or something like that, with a second layer to say "no crlf" ?

Nicholas Clark
-- 
Even better than the real thing:        http://nms-cgi.sourceforge.net/

Reply via email to