XDXF: Inconsistencies in transform

[RasmusKard]

Currently XDXF with CSS/JS transforms the keywords differently compared to the regular XDXF transformer.

With CSS/JS: Only first keyword is added into a <div>
With regular transform: Every keyword is added into a <span>
(With XSL=false my <ex> blocks also don't seem to be transformed, haven't looked into the code related to that to see what the issue is there)

Would a PR combining the two transforms for better consistency be welcome? I see that in #570 it was decided to separate them instead of using an read option.

If so, how should the keyword transform work? I'm expecting every key after the first to be used just for indexing not for display since if I'm not mistake that is the way to display inflections/forms. Should there also be an read option related to this?

[ilius]

@soshial