In an XML document, if you have an "input" tag that is not self-closing (the tag looks like "<input>...</input>" instead of XHTML default "<input/>"), you can't fold it. It's more of an annoyance than a real problem, but I've attached a screenshot to show this problem. In the screenshot, I'm editing an XSLT document in SciTE 1.73 (official build as of March 31, 2007) on Windows XP.
I looked through the source of the current HTML lexer, which is used to parse XML also, and found the problem, though I don't know enough about how the lexers in SciTE work to fix it. It's in the "classifyTagHTML" function, lines 239 - 247. From what I think is going on, the current tag is always compared against a list of self-closing tags, regardless if we're lexing HTML or XML. If the current tag is in the list, we don't fold it. Removing these lines causes input tags to fold in both XML and HTML.
The current behavior is correct for HTML, but it's slightly off for XML. (There is a comment that says the current check is only a quick hack, so this may be a known bug, but it wasn't in the bug list here. My apologies if this is a repeat.) From what I can see, fixing this would be as simple as figuring out if we're lexing XML or HTML, and then skipping the don't fold behavior for XML documents. However, from my limited tests, figuring out the current language is somewhat difficult. I'll keep looking into this myself, but I'm hoping it will be a rather simple fix for someone who understands how things work.