[xmlroff] New xmlroff, libfo-examples, xmlroff-testsuite, xmlroff-testing

[Tony G. <Ton...@Su...>]

Continuing the monthly releases for a record third month in a row, the
following are now available for your downloading pleasure:

 - xmlroff-0.2.7
 - libfo-examples-0.2.7
 - xmlroff-testsuite-0.2.7
 - xmlroff-testing-0.2.7

(There will also be spec-dump-0.3.1 once I manage to upload it.)

xmlroff-0.2.7 fixes a bug where an fo:block that was broken across a
page had the entire fo:block repeated on both pages.

It also adds support for resolving .length and .conditionality
components of padding-* properties, although padding-before is the
only one that's implemented in the area tree so far.

It also installs more headers than previously, plus it incorporates
the fix for segfaulting on unsupported properties that was provided by
Charles Bozeman.

The extra headers are installed because I had a day off this month and
found the time to write some examples (now in libfo-examples-0.2.7) of
programs that work with the FO and area trees:

 - xmlroff-text-tree demonstrates accessing the FO tree as FoNode
   objects.

 - xmlroff-text-tree2 demonstrates accessing the area tree as
   GObjects.

 - xmlroff-gmodule and gmodule-area-tree demonstrate using a loadable
   module that accesses the area tree as GObjects.  You need to run
   'make install' to install gmodule-area-tree for it to work.

The programs all just show a GTK+ GtkText widget that shows the nodes
in their respective trees.  It's a small step towards the tree browser
that I was talking about last year, and an even smaller step towards
the "area-tree adjust" phase in xmlroff processing, but it is a
start.  And all of you are welcome to continue from this start.

The next step (other than making modules for code that's repeated in
every example program) is to hide the GModule interface inside a
GTypeModule so the loadable modules can be treated as objects.

A further extension of that idea is to make the FoDoc subtypes into
loadable modules so you don't have to load a backend if you're not
going to use it, but that is a low priority for me.

There were only minor changes to the testsuite and test setup.
However, my tests are sometimes getting different (and bad) results
for the simple.fo test: I got the different result for a while, then I
didn't, and now it's back again.  Can someone check if they also get
different results for simple.fo with xmlroff-0.2.7?

My priority for June is to get past my fixation on padding properties.

I expect that my priority for July will be plugging memory leaks.

Regards,


Tony Graham
------------------------------------------------------------------------
Web Products, Technologies and Standards           Phone: +353 1 8199708
Sun Microsystems                                              x(70)19708
East Point Business Park, Dublin 3, Ireland