I'd like to use APR::Request's custom parser under some different
environments
like FastCGI, PerlEx and PersistentPerl. All these tie STDIN to some
package, which defines
methods to read a particular socket. APR::Request::CGI doesn't work,
because the
operating system's STDIN is not used.
Currently, to feed APR::Request's parser I have to do the following:
$length = $ENV{CONTENT_LENGTH};
read(STDIN, $buffer, $length);
[...]
$b = APR::Bucket->new($ba, $buffer);
[...]
This works nice for simple requests, but not good for file uploads,
especially not good for the previously mentioned environments,
because they are persistent just like mod_perl and filling $buffer to the
limit would take a lot of RAM, which is bad.
If I understand the "file bucket" conception right, if I supply a file
handle
instead of a scalar to APR::Bucket->new, it should work as desired,
because APR::Request's parser has a way to limit the number of bytes read.
APR's documentation says something about file buckets:
"It's important to know that certain bucket types (e.g. file bucket),
may perform a split and insert extra buckets following the current
one. Therefore never call C<L<$b-E<gt>remove|/C_remove_>>, before
calling C<$b-E<gt>read>, or you may lose data."
but it doesn't say how to create one.
Apache2::TieBucketBrigade won't work for me.
"Will Whittaker" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
>
> On Fri, Sep 16, 2005 at 12:25:24AM +0300, Nikolay Ananiev wrote:
> >
> >
> > I'd like to do the following:
> > open $fh, '<', 'some_file'; # A very big file
> > $b = APR::Bucket->new($ba, $fh);
> >
> > while($b->read($buffer, 2048)) {
> > .....
> > }
> >
> > or
> > $b = APR::Bucket->new($ba, \*STDIN);
> >
> >
> >
> >
> >
> >
> I'm not sure what you are trying to do in the above. If you are trying to
> make brigades usuable as filehandles then try Apache2::TieBucketBrigade
> http://search.cpan.org/~mock/Apache2-TieBucketBrigade-0.07/
>
> mock
>
