There has been a demand for annotation support which is ignored after validation. I have few comments regarding your proposed solution. You are storing references to the appinfo/documentation dom nodes. First, this means that you are keeping around the whole dom tree in memory. Second, once the schema is processed, it's dom tree representation is no longer available in memory (it's deleted). I think the solution would be to follow the same mechanism as the SchemaElementDecl class. Create a SchemaAnnotation, SchemaAppinfo and SchemaDocumentation classes that would be created and filled by TraverseSchema as it's processing the corresponding dom nodes.
Khaled Sir Woody Hackswell wrote: > On Wed, 2 Apr 2003, Gareth Reakes wrote: > > > Hi, > > we saw it :) We were just waiting for the most appropriate person > > to answer. I guess that has come down to me. There was some code written a > > while ago that did this. I did not make it into the last version. There is > > some demand for it and I would be happy to review/ commit the code if you > > are happy to put some work in writing some tests for the stuff. Does this > > sound OK to you? > > > > Gareth > > I was looking for some input mainly on the API I suggested. If what I have > is ok... needs more flexibility? It's crap... etc. If it's okay, I'll > clean up my code and get a patch ready. Should the patch be againt v2.2.0 > or against CVS? > > As for the tests... I'm not familiar with the test mechanisms of xerces. > Any pointers on where to look in the source tree to get a good example? > Thanks. > > -Richard Balint > > ----- > veni, vidi, vomi > > [EMAIL PROTECTED] http://sir.woody.hackswell.com > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
