Comments on: Fun with filters

By: talijanska

talijanska — Thu, 04 Dec 2003 05:20:24 +0000

olá otários de merda.

By: Chris Rimmer

Chris Rimmer — Fri, 24 Jan 2003 16:39:48 +0000

Unfortunately it also finds and replaces in links, image names etc. So if you fetch http://www.oracle.com and replace “ora” with “xxx” you end up with links to xxxcle.com and all the pictures disappear, because it is looking at xxxcle.com for them!

By: Yoz

Yoz — Sun, 01 Dec 2002 12:47:01 +0000

Earle: Cheers for the bug report; I’ve fixed that.
Paul: HTML::Munger’s nice but doesn’t actually do that much. Plus, I think that ignoring the fact that HTML and HTTP already have a (mostly) good method of changing the base URI is a bit silly. I may use it for a future filter rewrite.
In other news, that example bug I talk about for much of this entry seems to be gone; the filtered page now calls the correct stylesheet, without my having changed anything. Odd.

By: paul

paul — Thu, 28 Nov 2002 17:55:42 +0000

You also get problems if html tags are fiddled with (eg replace ‘<‘ with ‘ ‘, or ‘div’ with ‘img’ on any page).
In other news, the HTML::Munger module does a lot of the hard work for you. It avoids your problem by doing it the hard way, by rewriting all relevant urls in the page rather than using a base href.

By: Earle

Earle — Thu, 07 Nov 2002 03:37:50 +0000

Nice! But as you've pointed out, not without some issues. (Try, for example, replacing all instances of "the" on, say, http://downlode.org/blog.pl with "these".) Still, good one.