Heads Up and Questions

[Arved Sandstrom <Ar...@chebucto.ns.ca>]

I just want to let folks know that Wednesday evening (GMT-4) I'll be doing a 
whole mess of commits, and I want to explain what they are to anticipate any 
problems.

1) New files to support <fo:page-sequence-master>, 
<fo:single-page-master-reference>, <fo:repeatable-page-master-reference>, 
<fo:repeatable-page-master-alternatives> and 
<fo:conditional-page-master-reference>, and some supporting files;

2) Changes to PageSequence and LayoutMasterSet to support the above;

3) Changes to properties.xml, PropertyListBuilder and StandardElementMapping 
to support the above (there are a number of new properties);

4) DELETION of SequenceSpecification, SequenceSpecifier, 
SequenceSpecifierSingle, SequenceSpecifierRepeating and 
SequenceSpecifierAlternating, and all 4 PageMasterFactories in 'layout';

5) MODIFICATION of Driver, CommandLine and FOTreebuilder to support the use 
of SAX2 (XMLReader vice Parser, and Attributes vice AttributeList);

6) new examples folder and examples to illustrate the new FO's.

All of this will immediately render any existing FO file null and void 
unless you continue to use an older FOP or update your FO source. The 
upgrade to SAX2 should affect nothing.

There should be no other code that is affected. For example, 
initial-page-number stuff in PageSequence is untouched.

I'd appreciate it if any concerns or observations are raised before 
Wednesday (i.e. today or tomorrow).

Anticipated capabilities: first cut will not support "blank-or-not-blank" on 
<fo:conditional-page-master-reference>, per se, although the property will 
be read. And the "last" value for "page-position" on 
<fo:conditional-page-master-reference> is also not acted on yet. I haven't 
figured out the most elegant way of doing either, and suggestions will be 
welcome. :-)

Caveats: the upcoming mods represent a "clean-room" implementation of the 
pagination portion of the spec. That is, I looked at no other examples, on 
purpose. A large number of examples (which I'm finetuning now) are included, 
so as to indicate what the code supports. My interpretation is open to 
dispute, but the code is easy to change. As Sebastian Rahtz commented wrt 
his own code, probably nobody has a 100% grasp on exactly what the spec 
actually means. :-)

I'm pre-announcing this in particular so as to avoid any conflicting commits 
immediately prior to, or during, when I do all of mine. Thanks.

Arved Sandstrom