Re: same handler with different port
valerio crini schrieb: Sorry, i 'had typo errors. I managed in this way: VirtualHost *:12200 Location /rufus/query SetHandler perl-script PerlResponseHandler Dada::Rufus::Handler /Location /VitualHost Another typo here. VirtualHost *:12300 Location /rufus/query PerlSetVar Login true SetHandler perl-script PerlResponseHandler Dada::Rufus::Handler /Location /VitualHost Same typo once more. So i can pass different params only by changing port number: according to you is it the right way? Yes. In addition to this, you only need the following: Listen 12200 Listen 12300 PerlModule Dada::Rufus::Handler But you knew that. Michael Ludwig
Re: same handler with different port
valerio crini schrieb: VirtualHost *:12200 [...] /VitualHost Another typo here. VirtualHost *:12300 [...] /VitualHost Same typo once more. I can't manage finding where are typos, I'm very sorry. Now you can. By the way, beware the list does not set the Reply-To header to the list, so if you just hit Reply, as I do, the reply by default goes to the individual who has posted the message, not to the list, as it is probably intended to do. Michael Ludwig
What can a child_init do?
Is there anything a mod_perl2 child_init phase can do to call attention to a problem? I moved a block of code from a post_config handler to a new child_init handler, without thinking much about the niceties of the move. The code contained a couple of 'die' statements, which I trust would prevent an Apache startup if executed in a post config handler. Q1: WOULD 'DIE' IN A POST_CONFIG HANDLER ABORT AN APACHE2 STARTUP? In the child_init handler, an executed 'die' did nothing noticeable. Specifically the message did not end up in the error_log. In the mod_perl2 doc pages, child_init is described as being 'of type void', which another page says means that the return value doesn't matter. I will change the 'die' to a '$s-log_error', and return the nastiest- sounding Apache return code I can find, in hopes that some future Apache might notice it. Q2: IS THERE ANYTHING A CHILD_INIT PHASE CAN DO TO ABORT ITS CHILD PROCESS, AND THUS CALL ATTENTION TO A SERIOUS PROBLEM? It's amusing that the 'child_init' process seems to be an orphan in the Apache2 world... Thanks in advance, cmac www.animalhead.com