I've got a set of perl objects that need to capture the output of other objects. I accomplish this basically by saving STDOUT and then opening STDOUT on the write-end of a FIFO:

pipe READ_FIFO, WRITE_FIFO;
open STDOUT_SAVED, ">&STDOUT";
open STDOUT, ">&WRITE_FIFO";

component->display(); #displays the component to STDOUT->WRITE_FIFO
open STDOUT, ">&STDOUT_SAVED";
close WRITE_FIFO;

$content="";
while (<READ_FIFO>) {
    $content .= $_;
}
close READ_FIFO;
return $content;
[download]

The problem is that there appears to be a 4k (4096B) buffer limit. This solution works great with small components (text less than 4k), but for longer ones, the code hangs on the "component->display()". It basically fills up the buffer and waits for it to be cleared. But since I'm doing this all in one process, it never gets cleared :-)

Is there a way to increase the buffer size?

BTW, I'm not married to this approach, but it seemed the cheapest. I thought storing up a buffer of text in a FIFO was cheaper than forking a process. I also don't want to use a file--try to avoid the overhead and headaches involved with temporary files (especially accross multiple platforms).

Suggestions on other ways to catch the STDOUT to a variable are also welcome! TIA