Interesting problem...

[Michael Grabenstein <mgrabens@popd.isinet.com>]

    Ok here is one that is stumping me...

    I have written a subroutine that uses regular expressions to
parse XML. For a particularly large XML file when I write a simple
program and run it on the command line it takes 1.5 to 1.9 seconds.
The same XML file when parsed through the same library from a CGI
takes 15-20 seconds.

    I have run the CGI under an Apache web server and a Netscape web
server with the same basic results. Running the command line version
at roughly the same time on the same machine yields the time
difference stated above.

    The time stamps are basically the same locations for the command
line version and the CGI version and only show time actually spent
parsing. Watching these things run, it is evident that the times are
not wrong (ie. not a decimal point problem in the time calculation).

    I have printed out each tag encountered and printed the time
stamp at that tag and it shows a nice linear progression through all
the tags in both the command line version and the CGI version.
(IOW: my pattern is not getting caught in a recursive loop, etc...)

    I was prepared for the CGI to be slower than the command line
version, but not by a factor of 10.

    Any ideas ?

Thanks,
    Mike



--
kill -9'em All and let Root sort'em out
         --From Slashdot
Opinions, Flames, Irritations are solely mine.
Useful, Productive information... the company claims.


**Majordomo list services provided by PANIX <URL:http://www.panix.com>**
**To Unsubscribe, send "unsubscribe phl" to majordomo@lists.pm.org**