[scotia]

Hi Boss,

Since upgrading to Frodo - and therefore downloading a fresh copy of your script - an error I pointed out in (post 137) is back. I needed to add 'www' to the searchURL:

Code:

searchURL = 'http://www.weatherzone.com.au/search/'

at line 580 of your code.

This might be a larger problem for XBMC, but when using a proxy (which I am), the code behind calls such as:

Code:

urllib2.urlopen(req)

does not resubmit the query to redirected URLs (when the initial method is POST). At least that's my assumption.

I reconstructed your HTTP POST with cURL. Without a proxy it works:

Code:

[root@dev-01 ~]# curl -s -q -0 -k -d q=3122 -d t=3 --header "Host: www.weatherzone.com.au" --header "User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT" http://weatherzone.com.au/search/ | grep Hawthorn
                    <li><a href="/vic/melbourne/hawthorn">Hawthorn, VIC 3122</a></li>
<li><a href="/vic/melbourne/hawthorn-north">Hawthorn North, VIC 3122</a></li>
<li><a href="/vic/melbourne/hawthorn-west">Hawthorn West, VIC 3122</a></li>

With a proxy, and WITH 'www' prepended to the searchURL it works:

Code:

[root@dev-01 ~]# curl --location --proxy http://proxy.thismonkey.com:3128 -s -q -0 -k -d q=3122 -d t=3 --header "Host: www.weatherzone.com.au" --header "User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT" http://www.weatherzone.com.au/search/ | grep Hawthorn
                    <li><a href="/vic/melbourne/hawthorn">Hawthorn, VIC 3122</a></li>
<li><a href="/vic/melbourne/hawthorn-north">Hawthorn North, VIC 3122</a></li>
<li><a href="/vic/melbourne/hawthorn-west">Hawthorn West, VIC 3122</a></li>

However, with the original URL and with a proxy, it fails:

Code:

[root@dev-01 ~]# curl --location --proxy http://proxy.thismonkey.com:3128 -s -q -0 -k -d q=3122 -d t=3 --header "Host: www.weatherzone.com.au" --header "User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT" http://weatherzone.com.au/search/  | grep Hawthorn
<no output>

The man page for cURL explains it thusly:

...When curl follows a redirect and the request is not a plain GET
(for example POST or PUT), it will do the following request with
a GET if the HTTP response was 301, 302, or 303...

There's the issue, at least with cURL. The original POST becomes a GET. I'd bet that the same code is in use behind the scenes in XBMC. It'd be great if an XBMC developer could take a look.

Anyway, I guess I'm posting this for two reasons:
1) Is there any reason not to prefix the searchURL with 'www' since you're being redirected there anyway?
2) In case anyone else is having the same issue.

Thanks,
Scott