Whew viewing the source of an XHTML page in Firefox 1.0PR1, Ξ and ξ entities in the source are converted to &#Xi; and &#xi; in the View Source window.

Confirmed. Strange, even in a listing with other entities, like http://www.degraeve.com/reference/specialcharacters.php, only Ξ and ξ get the hash sign added.

It looks like the code that outputs view-source entities is outdated (like we didn't used to consume the # on hex entities, so we guessed). We now *do* do this. I'll attach a patch, but I need bug 263083 to be checked in first.

Don't prepend a # for no apparent reason.

Maybe I'm missing something, but shouldn't that !theStr.LowerCaseEqualsLiteral("xi") check ensure that we don't muck with ξ? What's the value of theStr coming throught this code for: "ξ", "e", "A" ?

These are some edge cases. Basically this is bz's list... View the source to test.

Comment on attachment 162594 [details] [diff] [review] patch v1 r+sr=bzbarsky if you just pass through GetStringValue() without assigning into theStr (I think that should be safe, but check).

Special casing "xi" is the wrong way to go. This bug also affects MathML named entities beginning with "x" -- and it would be ludicrous to also special case xwedge, xvee, xnis, xcap, xcup, xnis, xutri, xdtri, xlarr, xrarr, xharr, xlArr, xrArr, xhArr, xmap, xodot, xoplus, xotime, xuplus, xsqcup, Xscr, xscr, Xfr, xfr, Xopf, and xopf.

Robin, did you actually read the patch?

For reference, this is the patch that will be checked in once the tree reopens.

Checked in.