|
From: Bruce M. <br...@mc...> - 2003-08-14 04:23:46
|
All, Sorry about the large commits... What is this all about? Well the configuration handling in babeldoc is basically handled in three ways but the data semantics are all the same... There was the TieredConfigurationHandler, the PipelineStageConfig and the ScannerWorkConfig. All of these are tree like data structures and are representations of the xxx.ddd.ggg=value configuration properties. Now, this is bad, since it makes a centralized handling, understanding hard or impossible. Now all these commits for the last two days was to conflate all these three into one structure: IConfigData based structures. And these ConfigData structures are fully integrated with the IConfigInfo structures which describe the configuration options in babeldoc. This means that we can automate all kinds of things like validity checking, unified GUI handling, unified loading, traversal, etc, etc. So the upshot is that this configuration handling does not add much at the moment except that it rationalizes the codebase. Notice that I have done lots of checking on the Pipeline stages but less so on the Scanners. I am fairly confident that if one works, then the other should too - thats the beauty of a unified data handling/configuration codebase. regards, Bruce. |