On 02/10/2011 06:37 PM, Anca Luca wrote:
On 02/10/2011 06:27 PM, Thomas Mortagne wrote:
Here is a proposal to reduce all get...Block to 2 methods which support matching any Block.
And in the same spirit of XPath:
interface Block { enum TraversalStrategy { /** Search in parents **/ PARENT, /** Search in children **/ CHILD, /** Search recursively in children (and children of children etc...) **/ CHILDRECURSE, to replace with DESCENDANT,
/** Search in previous siblings **/ PREVIOUS, /** Search recursively in previous siblings (and parent previous sibling etc...) **/ PREVIOUSRECURSE,
I'm not sure what this means exactly, but if there is an equivalent in the xPath axes, we could use it. In any case, to find a name with one word, if there's not, I'm gonna start doubting that we need it :) or that anyone but us will understand what it does. I think it might be the PRECEDING xpath axis, but I am not sure.
/** Search in next siblings**/ NEXT, /** Search recursively in next siblings (and children etc...)**/ NEXTRECURSE
same as in the previous case. We could add ANCESTOR.
}
[...]
List<Block> getBlocks(BlockMatcher matcher, TraversalStrategy traversalStrategy, boolean recurse);
Block getFirstBlock(BlockMatcher matcher, TraversalStrategy traversalStrategy, boolean recurse); }
interface BlockMatcher { boolean match(Block block); }
et on refactor tous les autres get*Block basé sur ceux la avec des BlockMatcher prédéfinis (ClassBlockMatcher, MetaDataBlockMatcher, etc...).
Note: the main use case for this change is to support MetaData search.
WDYT (especially on the vocabulary like "TraversalStrategy") ?
XPath calls these "axes", http://www.w3.org/TR/xpath/#axes .
We could stick to the same naming, since it's well known in the tree document model world.
Otherwise I guess I agree with the change, although any generification and abstractization of the current already abstract model scares me.
Thanks, Anca
_______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs