[Booki-dev] Booki changes (bookizip branch)

[Douglas Bagnall]

I have been working on import and export in the bookizip branch.
(http://booki-dev.flossmanuals.net/git?p=booki.git;a=shortlog;h=ref/heads/bookizip)

The two main things I've worked on are:

1. Improving the metadata round-trip survival rate.  Now most if not
all archive.org metadata get through unscathed.  Booki contributors
are still not listed.

2. avoiding occasional clashes in identifiers.  There are a dozen or
so bug reports caused by this.

Along the way I've changed quite a few other things.  For example,
I've altered the export code to use an xml parser (lxml) rather than
regular expressions to fix up src attributes (it's actually quicker as
well as more robust), and I have shifted various hard-coded URLs into
configuration.

While I was doing that, Aco was spurting off in another direction
altogether, so there will be a bit of work in merging the two.  I
intend to make a 'bookizip-rome' branch upon which to reapply these
patches.

cheers,

Douglas