Skip Menu |

This queue is for tickets about the CGI-WML CPAN distribution.

Report information
The Basics
Id: 4123
Status: new
Priority: 0/
Queue: CGI-WML
People
Owner: Nobody in particular
Requestors: thomas.hoppe
Cc:
AdminCc:
Bug Information
Severity: Important
Broken in: (no value)
Fixed in: (no value)

History Show all quoted text
  Thu Oct 16 05:41:20 2003 Guest - Ticket created
Subject: -cache_control will never work as described
Hi Andy, System Spec: This is perl, v5.8.1 built for i686-linux Linux login.info-miner.de 2.4.20-019stab031.3.777-smp #1 SMP Mon Sep 1 21:41:58 MSD 2003 i686 unknown I was again experimenting with CGI::WML. My problem: I want to cache the generated output on the mobile. Which does not yet happens in my application, on every prev/ the script will be invoked once again, doubling the amount of kB which need to be transfered, making the entire app quite expensive, For that purpose, I was trying to write out the Cache-control information in the way you decribed it. But that will never work, since the allowed header information is tested/filtered in CGI. And Cache-control does not belong to the allowed header information, so no output is generated in that header line :( Maybe I will find the time to invent a solution, but I am not sure. By the way I have solved my problem with wml_to_wmlc. I just write first into a buffer and compile than. Regards, Thomas
  Thu Oct 16 08:12:35 2003 Guest - Correspondence added
From: thomas.hoppe [...] berlin.de
[guest - Thu Oct 16 05:41:20 2003]: Show quoted text
> Hi Andy, > > System Spec: > This is perl, v5.8.1 built for i686-linux > Linux login.info-miner.de 2.4.20-019stab031.3.777-smp #1 SMP Mon Sep 1 > 21:41:58 MSD 2003 i686 unknown > > I was again experimenting with CGI::WML. My problem: I want to cache > the generated output on the mobile. Which does not yet happens in > my application, on every prev/ the script will be invoked once > again, doubling the amount of kB which need to be transfered, > making the entire app quite expensive, > > For that purpose, I was trying to write out the Cache-control > information in the way you decribed it. But that will never work, > since the allowed header information is tested/filtered in CGI. And > Cache-control does not belong to the allowed header information, so > no output is generated in that header line :( > > Maybe I will find the time to invent a solution, but I am not sure. > > By the way I have solved my problem with wml_to_wmlc. I just write > first into a buffer and compile than. > > Regards, > > Thomas
  Thu Oct 16 08:18:33 2003 Guest - Correspondence added
From: thomas.hoppe [...] berlin.de
[guest - Thu Oct 16 05:41:20 2003]: Show quoted text
> For that purpose, I was trying to write out the Cache-control > information in the way you decribed it. But that will never work, > since the allowed header information is tested/filtered in CGI. And > Cache-control does not belong to the allowed header information, so > no output is generated in that header line :(
No that wasn't the reason, it was just a sloppy analysis :) Show quoted text
> Maybe I will find the time to invent a solution, but I am not sure.
Cache-control didn't worked be cause rearrange was build for HTML ! See comment in CGI.pm. In the header routine of CGI::WML you have to use CGI's unescapeHTML as follows to fix this bug: foreach (@leftover) { next unless my($header,$value) = /([^\s=]+)=\"?([^\"]+)\"?/; $leftover{$header} = $self->SUPER::unescapeHTML($value); } Regards Thomas