[Bug-wget] Convert-links and 302 redirect

[Pavel V.]

Hi !

I'm trying to download a some site using Wget.

All site pages has a link to main page ("/") on its logo.
One of the site pages has a link (for example, "/news-302/") which results a 302
redirect to "/" (to main page). The Wget follows that redirect and result is
downloaded as a file "news-302/index.html". That is correct behaviour.

But later, when Wget does links convert, an problem occurs: all pages has "/"
link replaced to "news-302/index.html" instead of expected "index.html".

The issue is reproduced in "GNU Wget 1.13.4", "GNU Wget 1.16" from Debian 
packages.

I have prepared a simple testcase of three pages.
You can reproduce this issue by mirroring http://areainter.net/wget/ by running

wget --recursive --level=10 --convert-links --page-requisites 
http://areainter.net/wget/

The obtained mirror is also accessible by http://areainter.net/wget-result/ url.

How to avoid this problem of wrong links conversion? Maybe that is a bug in 
Wget code?

-- 
Regards,
 Pavel                          mailto:address@hidden