Okay, here we go, how to debug svggraph-webif:
First, your browser has to be able to do SVG images! Here's a page with some samples, if you don't see something like the little icons when you click the links your browser probably doesn't support SVG:
http://www.croczilla.com/svg/samples/
If those do show up, then let's see if you can load these svg documents! From here on, replace "http://openwrt/" with the correct value to be pointing at your openwrt box.
http://openwrt/svggraph/graph_cpu.svg
http://openwrt/svggraph/graph_if.svg
Neither of these will "work" now, but if you get an empty graph that says it can't load data, then your browser loads the svg images for this package properly. If you get a download prompt, then your browser is confused or (if you use Firefox 1.5+) then your web server is misconfigured. You'll need this line in /etc/httpd.conf (set automatically by version 0.3 and above):
These SVGs make AJAX-like calls to the openwrt box to get the data to put in the graph, located at these URLs:
http://openwrt/cgi-bin/webif/svggraph/data.sh
http://openwrt/cgi-bin/webif/svggraph/data.sh?if=br0
Where the first link above should display CPU data, like this:
Mon Feb 13 10:10:55 EST 2006
cpu 146970 0 1145264 62863984
And the second will display interface traffic. Every different piece of hardware (and configuration) can have different interfaces, so you may have to choose an interface name that actually exists on your system:
Mon Feb 13 10:11:14 EST 2006
br0:3602758287 8927578 0 0 0 0 0 0 1264113438 9404396 0 0 0 0 0 0
In other words, for cpu data, the script expects the date on one line, followed by the word cpu, then four separate numbers, on the next line. For interface data, again the first line is the date, the second line is "interfacename": followed by 8 numbers, and 8 more numbers. The script is concerned with the first number from each of those groups. (If you "cat /proc/net/dev" you'll see that they are the number of bytes transmitted and recieved.)
If things are not working for you still, please follow these instructions:
First, install Fiddler (Or something similar if you aren't on Windows). Then configure your browser to load pages through its proxy, and visit each of the URLs listed here once, or visit the page that gives you a problem. Then for each URL, select it in the list on the left, click Edit->Copy->Session, and paste that into a text file. Then, email it to me (link on the left of each post) and I'll get back to you! This way I'll have a lot of the relevant information and we won't have to play email/forum tag to get the details.