output flush not working

Anon

I have this code snippet:

ob_implicit_flush();

$fp = popen ( \"/perl/test.pl\", \'r\' );
while ( $results = fgets ( $fp, 1024 ) ) {
echo $results;
}
pclose ( $fp );

ob_end_flush();

the test.pl program being executed is simply printing a line and sleeping:

for ( $i = 1; $i <= 5; $i++ ) {
print \"line $i<BR>\";
sleep 3;
}

Yet the line by line flush is not happening. All of the output is displayed at once when $fp is closed. Anyone know why output flush is not working here?
Thanks in advance!

van__

Hey Jill, nice to see that you saw my response to your last post.

The problem here is this...

while ( $results = fgets ( $fp, 1024 ) ) {
echo $results;

}

the fgets counteracts your intentions for the while loop though fgets returns true, fgets keeps on fgetting before while has a chance to echo.

Unfortunately, you're using a perl script, and I don't have time to write two scripts to test (I'm supposed to be debuggin some production code right about now)

But, try using a passthru();
instead of a filesystem call.
The passthrough will execute the perl script and output the results directly to the browser.

Anon

Thanks for all your help, Van.

Surprisingly, the passthru didn't work. I couldn't use it if it did anyway, because I need to check each line of output (and possibly change it) before putting it to the browser.

Your 'for' loop worked great with the output buffering off, it even worked without turning off buffering...

My dilemma is executing an external program and getting a line of output at a time from it 'massaged' and printed on the browser as the output comes in (without waiting for all output). hmmm... if anyone has any other ideas, please help!

van__

Well, your problem starts to get tricky.

I noticed in the perl script that you are using sleep(), I assume that's becuase the data fed back from your real perl script for production will be constantly updating...

and if that's the case, that is a fun little puzzle. If you don't need a constant update from a perl script, but just want to check output and write out a time seperated line,
then the answer is pretty simple.

Let me know what you need get here.

Anon

Van, thanks again for offering your help!

The sleep was just a test program to see if this whole thing would work. In reality, here is what I am trying to do:

PHP script will receive a list of website URLs to hit (from user choices on a form)
it passes the URLs to a perl script that fetches the URLs in parallel
the perl script parses the output and gives a summary of data about the website pages as each one responds (hence delayed lines of output)
PHP receives each summary and puts it into a <TABLE> for the browser

If one of the websites failed to respond, I need to put those results last, therefore creating the need to look at these results before output to the browser...

What are your ideas? Thanks!!!

van__

Jill,
It's probably best to recreate the perl functionality with PHP.

cUrl is a nifty utility for remote URL connections and allows you to GET/POST data as well as send/recieve cookies. I use this all the time to connect to remote site and parse them, plus it gives nice responses and allows you to read the server response headers.

If you must use the perl script, it's possible that by doing some tweaking, you could get the script to work better with your PHP script.
One thing you can do is have PHP call the perl script multiple times, one for each page request, it can then take each request result and parse it before handing it to the user.

But regardless of what direction is taken, the end result needs to be one where PHP can grab each result, parse and output before starting the process over again.

This thread has already fallen off the page,
so feel free to email me and we can go over it via chat or something.