PHP Program Randomly Stops

[Charlie-Boo]

Who knows how to do any of the following (1-3)?

1. "You're going to want to run it not through the web server, but via
the command line."

2. "You can use a combination of PHP with client side code to keep the
script running. Set a particular number of pages to be processed, a
number lower than when you experience the timeout. As long as there
are pages to be processed print a client side script, e.g. JavaScript,
to reload the page. This will reload until all pages are done - then
don't output the reloader code."

3. "You are running the script from a browser. The browser receives
output from your PHP script - or waits for it. Send output to the
browser that triggers the reloading of the page. This makes sense when
you invoke (whatever snoopy is) the class a finite number of times,
e.g. 100 URLs at a time. Once the URLs are processed keep track of
the number of processed URLs via a session variable. Then output some
JavaScript to the browser that reloads the page. Here's an example
How do I "print" a client side script?{: print
"<script>document.location.reload();</script>"; Alternatively you can
do it in PHP if you make sure there is no output sent before
redirecting. You could use the header() function to redirect the
script to itself. header("Location: ".$_SERVER['PHP_SELF']);"

Explanation:

I have written a PHP program to retrieve the HTML of web pages, using
class SNOOPY http://sourceforge.net/projects/snoopy to translate the
URL into the HTML. After about 100-300 references to snoopy, which
takes about 5-15 minutes, it stops.

1. I start it using Internet Explorer.
2. There is no error message.
3. The program has no reference to set_time_limit.
4. max_executiontime=0 in php.ini .
5. A program that simply executes an endless for, periodically
printing out the current time, runs indefinitely, which I manually
stopped after 2 hours.
6. I am running it on my single PC window 98 webserver Apache from
PHPTriad http://sourceforge.net/projects/phptriad .

I received the above three answers, but I don't know how to do those
things. I program only in PHP (and HTML.) I couldn't get any further
explanation.

Charlie ch***@aol.com

[Andy Hassall]

On 29 May 2004 12:45:11 -0700, ch***@aol.com (Charlie-Boo) wrote:

I have written a PHP program to retrieve the HTML of web pages, using
class SNOOPY http://sourceforge.net/projects/snoopy to translate the
URL into the HTML. After about 100-300 references to snoopy, which
takes about 5-15 minutes, it stops.
OK, if you have a script that runs for that long, it's inappropriate for it to
be run within a single HTTP request.
Who knows how to do any of the following (1-3)?

1. "You're going to want to run it not through the web server, but via
the command line."
http://uk.php.net/manual/en/features.commandline.php

For a script that takes 15 minutes, this by far the preferred method.
2. "You can use a combination of PHP with client side code to keep the
script running. Set a particular number of pages to be processed, a
number lower than when you experience the timeout. As long as there
are pages to be processed print a client side script, e.g. JavaScript,
to reload the page. This will reload until all pages are done - then
don't output the reloader code."
Ugh.
3. "You are running the script from a browser. The browser receives
output from your PHP script - or waits for it. Send output to the
browser that triggers the reloading of the page. This makes sense when
you invoke (whatever snoopy is) the class a finite number of times,
e.g. 100 URLs at a time. Once the URLs are processed keep track of
the number of processed URLs via a session variable. Then output some
JavaScript to the browser that reloads the page. Here's an example
How do I "print" a client side script?{: print
"<script>document.location.reload();</script>"; Alternatively you can
do it in PHP if you make sure there is no output sent before
redirecting. You could use the header() function to redirect the
script to itself. header("Location: ".$_SERVER['PHP_SELF']);"

Sounds like a perversion of what HTTP is meant for, but possibly workable as a
last resort. Unnecessarily complicated though; just go for (1).

--
Andy Hassall <an**@andyh.co.uk> / Space: disk usage analysis tool
http://www.andyh.co.uk / http://www.andyhsoftware.co.uk/space