Re: Enfiladealigner (was: Re: [Gzz] Ids)

[Benja Fallenstein]

Tuomas Lukka wrote:

> > But if it's not a VStreamTexter, what do we do? For example, if it's a 
> > cell texter handling included objects through containment, as you 
> > proposed (cannot be represented in an Enfilade1D).
> >    
>
> Well, that's the strategy's problem! You can have in gzz.fuzzy a factory
> that creates the right kind of strategy for each celltexter or something.
Ok.

> > Of course. The question is how? StringSearchers don't want to know about 
> > the deltaspace (they're completely independent of the gzz package) and 
> > besides, the deltaspace is not in use yet. (Is someone going to bring it 
> > into use?)
> >    
>
> StringSearchers could IMHO know about Delta objects, specifically 
> ContentDeltas. Then it would just be a matter of a generic hook mechanism.
Then they would be limited to cell content searchers-- the current 
interface is more general, and I like that. I think there should be some 
intermediate class that does the hook's job-- maybe simply the factory 
you propose above.

> Of course, one point I'll be adamant about: the stringsearcher hook
> MUST be run *after* the animation: it's too slow as it is.
It's too slow as it is because it re-builds the *whole* StringSearcher. 
But of course it is correct to do it after the animation (except that 
pending updates *must* be done before user searches are processed, 
because we don't want to exhibit inconsistent behaviour).

- Benja