Not willing to wait until Christmas, I decided to look at the Ha'aretz RSS problem. My diagnosis of the problem: something is wrong with the Ha'aretz feeds, but I am not sure what it is.
The site's layout does not seem to have changed significantly. Also, the code returns empty feeds when run on this server, but it works perfectly on the FAS server. Basically, when it is run here, my code refuses to return all of the links on the relevant Ha'aretz pages (including the crucial article URLs). If I could pinpoint why this was happening, I could fix it. However, I am still not sure which links it is returning; I initially thought it was only finding the first thirteen or so, but further experimentation seem to suggest that there is a section in the middle of the page (which, entertainly enough, includes the literal middle of the page - where the articles are displayed) whose links are disappearing into the ether.
Since I have spent almost two hours attempting to debug this (with no end in sight), I have temporarily removed the Ha'aretz feeds. When I wake up later today, I will take another look at the problem.
UPDATE (22 Dec 2004, 21:15)
I am still not entirely sure what the problem is. The only semi-rational explanation that I can come up with is that the complexity of the table-based layout of Ha'aretz's HTML pages are causing HTML::TokeParser to skip some tags. The FAS server has a slightly newer version of Perl than this host, so I am going to try switching to a local version of TokeParser.
UPDATE (22 Dec 2004, 21:40)