Have discovered lots of things about guidelines.xsl, but no closer to solving this problem. I should mention the one bit where I am not sure I agree w/ you on this, though, Hugh:
we may be better off just removing @source='UND' from the cases where it causes an error for now
I'm wondering if we have to make a temporary kludge fix, it might be better to remove the test from the epub build, since we don't test for it anywhere else. (Yes, we may end up w/ bad links in the HTML, but we may already, the links aren't helpful in this case anyway, and at least the HTML and the ePub would be more likely to match.) Of course, that said, I don't know *how* to remove the test in the epub build, as it is in java. Do you (Hugh) have time tomorrow to Skype (or GH or whatever) on this?
I concur with most of this: * does not seem to have anything to do with UND itself - may come up w/ UND because there are so many of 'em * probably occurs elsewhere, just not getting caught
I am in the middle of ascertaining the paths to all of the <egXML>s in BACKLIST, in the hopes that if I locate the 8 problematic ones in there, something will jump out. Maybe we should talk tomorrow?