[Bug-wget] (no subject)

[Jan Peter Leentfaar]

Hi.

I recently used wget to retrieve a page from my own webserver like

wget -P ~/wget/ --recursive --convert-links  http://myserver/dir/page.html
(page, dir and server names changed)

In this page, I have an iframe with 
src="/another_dir/another_page.html?parameter1=value1&parameter2=value2"

This is explicitly done this way to be able to call a function in the iframe 
from the main page.

However, the iframe source is converted to 
src="http://myserver/another_dir/another_page.html?parameter1=value1&parameter2=value2";
In addition, there are files in ~/wget/myserver/another_dir/ with the filename 
including the parameters. I have:

~/wget/myserver/another_dir/another_page.html
~/wget/myserver/another_dir/another_page.html?parameter1=value1&parameter2=value2

Since the main page now (not as before) has an absolute src pointing to my web 
server and not to the local files, both of these are ignored anyway.

Now this causes the local page.html and the iframe to be on a different domain, 
from the browser's point of view. Therefore, calling the function in the local 
page after retrieval, causes an error.

Somehow, I would have expected the frame source being converted to a source 
relative to page.html (which) with the part after the question mark unchanged, 
and no extra another_page.html?...-files like all other references (like real 
links and .js and .css references in the document head). 

Did I miss a combination of parameters with which I can get back what I 
expected (i.e. converted relative iframe paths) or is this unexpected or 
undesired behaviour?


I would not like to change the urls from /another_dir/xxx to ../another_dir/xxx 
because that would not allow me to simply move the main page to another 
directory level without changing the references as well.



Greetingz,
J!P