On Tue, Jan 02, 2001 at 03:40:50PM +0000, Stephane Bailliez wrote: > > I don't like the idea of being so tighted to a specific implementation > (especially a xml parser) but AFAIK this feature unfortunately does not > exist in other parsers (except with the DOM L2 importNode), and I didn't > feel like cloning all elements of the tree right now (this is prototype > after all, no need to reinvent the wheel :-). > > Once this is validated I would prefer to remains parser independant and add > additional code, at least it will avoid parser transition problem, and plus > performance should not be an issue. > > > From: Conor MacNeill [mailto:[EMAIL PROTECTED] > > > > Sure, I didn't consider antidote as I don't use the GUI, at > > least not yet. > > Not that keen on tying the GUI or JUnit to a particular XML > > parser either. > >
Antidote's implementation specific dependence on the Sun implementaion is temporary. It's a quick hack to get the parsing behavior I want while we are all waiting to see what Ant 2.0 ends up looking like. I don't want to waste time on parsing support until I see if I'm going to have to mirror the data model or not. Regardless, I don't recommend bundling Antidote in any installation package as it's not ready for prime-time. That being the case, if the ant core works with xerces, then I'd publish that compatibility. sim -- Mustard Seed Software mailto:[EMAIL PROTECTED]
