Front page | perl.perl5.porters |
Postings from May 2003
Nicholas Clark <nick@ccl4.org>:
> > a) use setvbuf() to discard the buffer, perhaps temporarily, or
>
> My understanding is that one never was allowed to call setvbuf() midway on a
> stream:
Right, thanks. Please forget about the 'perhaps temporarily' part. I
added this as an afterthought in my message anyway. The essential
problem is that I would like to do
$rec = <FH>;
but that implies some sort of buffering, and the buffering confounds
concurrent access. So I planned to use setvbuf to disabled the
buffering while stil permitting use of <FH>.
Kurt Starsinic <kstar@cpan.org>:
> How about mmap(),
I need features that are provided by standard perl, preferably back to
5.005. As far as I know there's no entry to mmap except through
nonstandard CPAN modules. PerlIO uses it internally, if it's
available, but doesn't make it available to me.
Kurt Starsinic <kstar@cpan.org>:
> or closing and re-opening the read() filehandle?
Not an option, I'm afraid. It may not be re-openable.
Using
open $newfh, "<&=".fileno($oldfh)
seems not to work for some reason that I haven't investigated yet.
Graham Barr <gbarr@pobox.com>:
> You could try using seek.
That works on my platform, but I worry that it's not portable. I
can't think of any reason why 'seek' should necessarily discard the
buffered data. Perhaps I'm overlooking something?