Bug: 2.3.0-prior: Oh html standards, how vB developers loathe thee...

[Paul]

vBulletin 2 fails to properly use "&" instead of "&" This happens all over the code. Everywhere. It's invalid. It sucks. Someone was lazy. I've complained before, but this was not going to be fixed.

Up until today, I thought this only affected the validity of the doctype, however, as I post the pronounciation of a word in my message, I note that the "&lt" has converted itself into a "<"

How can this be fixed? (as if 2.3.x, the officially supported product here will ever see another bug-fix release in our lifetime). Convert the "&" symbol, which per HTML standards should never appear by itself, into "&amp;", the proper way of displaying an ampersand, before displaying posts.

I won't bother posting a fix here. The bug tracker doesn't accept entries for vB2. Apparently you don't care, so why the effort on my part?

Anyone else seeing this quirk now knows why.

[Paul]

This raises the question whether or not vbulletin users want to recognize &<stuff>; as html markup or as valid ways to post information. When posts are htmlspecialchars()'ed, non-standard things are converted into their ascii codes. Converting the & will cause things like "…" to start appearing in posts. I'm guessing this is why the amp conversion was removed at some point.

A new option could be created to allow these symbols in posts--in that case only convert the ampersand when the symbol isn't valid (it should start with a & and end with a semi-colon). It appears that IE will render codes that neglect to end with a semi-colon.. perhaps this is part of the quirks mode? For now, I'm adding a regex that checks to ensure the codes are valid before rendering them appropriately to allow codes posted by international users to appear in messages.

[Paul]

I've chosen to use the following:

$bbcode=preg_replace("/&(?![#a-z0-9]+;)/i","&amp;",$bbcode);