I'm of the personal opinion that you decode either zero or one times. It doesn't help for RSS, but the times that I've written my own schemas, I've used two separate types -- the normal 'string', which I didn't decode at all, and a type 'embedded_xml', which was decoded once, and only once.
In dealing with CGI programming for the last 10 years or so, I've lost count of how many of the early cross site scripting flaws were from people using multiple-pass URI encoding, or multiple pass HTML encoding. (or both ... but technically, a single HTML encoded URI encoded URI is legal ... I use it for mailto links all the time)
In reply to Re^4: HTML from single, double and triple encoded entities in RSS documents
by jhourcle
in thread HTML from single, double and triple encoded entities in RSS documents
by Anonymous Monk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |