* Fix content extraction and html output
- Content extraction was breaking in case it included a templating
instruction
- the html output was non-compliant html by having nested <ul>...</ul>
not belonging to a <li>...</li> element.
* refactor processListsAndLines in smaller functions
* fix date consistency
- title was using a date reconstructed from UTC time which led to
discrepancy between the content date and the title
- pubDate was based off the gmt time at which the process was
run. Instead, if the last possible entry time (23.59.59 gmt-12) is
passed, we use that, if not we use current time. This should reflect
more or less a form of 'publish' date status from this portal.
* ensure empty-ish page doesn't lead to entry creation