[reportlab-users] Unicode, subversion and Version 2.0

Andy Robinson reportlab-users@reportlab.com
Mon, 21 Jun 2004 10:15:06 +0100

3 important notices...

Subversion Switchover
Here is a very belated update on our subversion switch,
which we originally wanted to do in May.  (For those who
missed the chat, we are moving to a new version control

It turned out to be quite hard creating a tool to move across
our repositories and to configure it right, and we found a few
security holes we needed to plug.  Took about five
weeks.  We then decided we should develop against it for a
few weeks as "User Acceptance Testing" before making an irrevocable
switch, which we are now doing.  It looks good so far although
it does have its quirks.  We're doing some work on our SVN
reportlab repository now, without the added burden of external users
but you'll see the cumulative changes soon in SVN or CVS.

If all goes well we will publish access details in the first few
days of July, and mark the CVS tree as "dormant" (i.e disable all
SF checkins, maybe rename __init__.py so it's unusable, and add a large
readme saying where it moved).  In the next ten days have a customer
deadline so it has to wait until after then.

Overall it's clearly a nicer system.

We had to get some emergency Unicode patches in place for a customer
last month (basically, accepting UTF8 in RML and auto-converting to
the relevant font on output).  We managed it, but it's hacky with
suboptimal performance.  Having started this, there is basically
no going back; some old (and arguably wrong) behaviours have to be broken
and we have to move to Unicode internally in platypus and RML.  So, I am
going to treat this as "the road to Version 2.0".  

When the new repository is publicly up, the first thing we will do is
make a 'stable branch', with a 1.19.x named release, and attempt to
maintain bug fixes on this for a while.

The trunk will then be a 'pre version 2', which currently runs but
may bite you on things like bullet characters from time to time.

Version 2.0
This fixes the scope nicely for Version 2.0.  We also have a joint
project on with a local university this summer on a better system
for authoring documentation, so the two driving forces will be
"unicode" and "better documentation".  Rough goal is "this summer".

In scope:
	- Unicode internally where possible
      - clean up font registration, families and caching
      - remove old xmllib-based paraparser and ensure that 
      - lots more test cases
      - defined what's the public API and what's private
      - new system for documentation authoring
      - better packaging and deployment

Deferred (i.e. "version 3.0":
      - changes to layout or paragraph algorithm
      - merge of graphics and platypus

Things which may change:
      - anyone doing very specific hackery with

That's my rough plan.  Comments welcome.  Volunteers welcome too
(after July 1st).  Take it with a large pinch of salt as usual and 
disregard any target dates until you see it happening ;-)

- Andy

Andy Robinson
CEO/Chief Architect
ReportLab Europe Ltd.
mobile +44-7976-355742
office +44-20-8544-8049