On Nov 21, 2003, at 11:08 PM, Robert Brown wrote:
If the only real purpose for the fork is so that the message can be displayed, why don't you just display the message in a form, and then go do the long operation single threaded. When the long operation finishes, continue to output to the same page and use some javascript to change the contents of the form field where it says please wait.
I wrote a web based chat over a year ago that makes use of the fact that if you specify a large value in the file size header, the browser will stay connected for quite a while. I can output posted messages to the browser and it will display them as they come. It will just patiently wait there with its busy indicator showing, but it works fine. If you experience timeout problems, just output a blank space character every 3 seconds or so. This will not change anything, as the text is being set via html rendering, sso spaces don't count.
If, on the other hand, you really need to fork a background task, you may be thwarted by the apache perl cgi module that keeps the perl interpreter in apache instead of forking a new perl for every cgi script. I do not know if you can fork a background task with perl-cgi-mod or not.
-- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]