so I went around and searched for the source code:
http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/LabeledSectionTransclusion/lst.php?view=markup&pathrev=109709And then read up a bit on how the parser works:
https://www.mediawiki.org/wiki/Manual:Parser.phphttps://doc.wikimedia.org/mediawiki-core/master/php/classParser.htmlSo from the looks of it, it sets up some hooks in the parser function in the setup function, so that whenever a <section> tag is found it is ignored (via the "noop" function which returns empty string) and whenever an lst function is found it is processed by "pfuncIncludeObj" which, using function "setupPfunc12" extracts the title of the target page from $frame and puts it in $title and the title of the start (and end, if it exists and is different, meaning that you could use #lst:target|startsection|endsection}} if you wanted to start from one section's start tag and end in another's end tag) sections from $args and puts it in $begin and $end. It also sets up some other helpful vars. Then it starts from the root of the DOM of the target page (received using "Parser::getTemplateDom" in the setup function) and starts going through the tree, looking for the target start section tag
without any special consideration for templates (that I can see. I have no idea what the splitExt function does). Once found, it starts copying all elements after it,
including the text from any templates it finds by using "Parser::replaceVariables" in "expandSectionNode", until it finds the appropriate end tag, after which it continues searching for the next start tag.
I'm assuming the problem is that Parser::getTemplateDom doesn't expand templates inside the target page and because the function doesn't expand templates when searching for the begin tag it never finds the start tag and thus never starts copying text. That or the parser somehow "sanitizes" the templates from tags it doesn't like, which includes things like "section" tags, because of the noop hook. Quick way to test it is to put a begin tag outside a template and an end tag inside a template. If it finds the end tag, then it's the first problem. If it doesn't it's the second problem.
Maybe I'll set up a wikia server at home and try to step through it with a debugger (when I'm no longer supposed to be busy with other stuff) so that I can figure out how it works exactly. Mostly out of curiosity. Maybe I'll even contribute a bugfix if it's easy to fix.