Action Item: Test RNG Refactoring

[Eliot Kimber <ekimber@contrext.com>]

I have made  a first stab at the RNG refactoring I outlined some time ago. I have pushed the results to the branch feature/rng-refactor, reflecting the current state of the DITA-2.0 branch.

As part of this I corrected some minor errors in the media domain RNG module: it uses "global-atts" (which has been reliminated) and explicitly declares @outputclass where it should not (@outputclass is now part of univ-atts).

I also corrected an explicit @outputclass declaration in the new linktext RNG declarations (part of common elements).

The refactoring moves the references to commonElements, metaDecl, tblDecl modules out of the topicMod and mapMod modules and into the document type shells. It also renames commonElementsMod to commonElementsDeclMo and renames tblDecl to tableDecl, making these modules' names consistent with our general naming conventions.

I tested the refactoring using the test docs I use for testing the RNG converter (I added a table to the topic sample to make sure table markup was OK).

I tried running these through the existing RNG-to-DTD converter, which ran but didn't produce the correct DTDs because the current code has special cases for commonElements and some other things. I would need to make the DITA 1.x code for these special cases conditional. I had started to do this in the context of the media domain but didn't complete it when I set my DITA work aside early last year.

ïI will attempt to get the RNG-to-DTD converter updated to handle the current 2.0 grammars and this refactoring as I can.

Cheers,

E.

--
Eliot Kimber
http://contrext.com