Fix problems in jmxml.py
While writing test code, found need to read in entry data from xml. Using jmxml.parse_entry() had a number of quirks, in part due to it being one of the first pieces of code written for jmdictdb. 1) Need to read KWs data and poke it into jdb module. 2) Need to mkxml() and poke the results into jmxml module. 3) Entr objects returned by parse_entry() have all the order/links fields (eg, rdng.rdng, sens.sens, gloss.sens, gloss.gloss, etc) set. Jmdictdb convention for entry objects an constituents leave these all set to None since their value is implicit is the position of the object in its containing list. They are usually only set just before writing to database. Having them None would simplify test code that currently has to change these values when position of object in list changes. Should either add option to jmxml to provide Entr objects without these fields set, or make non-setting always and modify jmparse.py (and whatever else) to call jdb.setkeys() before using the Entr objects in the numbering is needed. 4) I don't think (not sure) the jmxml will parse my extended jmdictdb xml format?
[Imported from JMdictDB Issues Tracker: IS-217/msg540]