I'm just shooting in the dark that this will be useful to someone that reads it; if anyone has a better idea where to send it let me know, or if this is a bug that has been fixed in later versions.

I am using Konq in KDE 4.2.2, and I had a slight rendering problem on the Unicode test page at http://www.alanwood.net/unicode/arrows.html . For the symbol with hex value 21AE it displayed instead the symbol for 21A, then an E and a semicolon. Similarly, for the symbol with value 21DE it displayed the symbol for 21D, then an E, then a semicolon. Other than than all the decimal and hex versions of the symbols were the same (as they should be). I viewed the source, which looked fine, and when I browsed away from the page and back everything appeared correctly. I have seen similar things on this site before -- once with ≨, which displayed a symbol that was not , followed by 268;.

So this bug is not easily reproducible, even for the person reporting it. As a programmer, I hate that kind of thing, but I thought I should throw it out somewhere (the KDE bug system doesn't want me to submit a bug from such an old version, which I respect). I suspect that it might have to do with the page being rendered after only being loaded partially. That is, rendering after the last data received was "&#x21A", then getting the next chunk of data, which is a perfectly normal "E;", and not going back and changing the "&#x21A". If that's the case someone that knows how the code works could probably set up a sandbox test for that sort of thing.