Store matching RegExps

[blake-regalia]

What is the expectation here for the underlying string that the regexps are being matched on? The reasonable thing that comes to mind would be to clarify in the spec that implementations must match the regex against some canonical string for a term such as N-Triple, but this opens up questions. Are blank node labels be preserved? Should the default graph be ignored if a RegExp is used?

[jacoscaz]

As Store implementations might not share the same underlying serialization format, I would push towards filtering constructs (#123) through which each Store implementation could build its queries rather than RegExp-based pattern matching (which only applies to strings).

[bergos]

I would propose to remove RegExp from the interface. @jacoscaz proposed some extensions in #123 . We could move the ideas from the other issue and the RegExp stuff into a advanded filtering and/or let implementors move ahead with a proposal.

[rubensworks]

I would propose to remove RegExp from the interface.

I strongly agree with this.

The RexExp params complicate the implementation of Store too much. While it may be useful for consumers, it does not really belong there IMO.

[RubenVerborgh]

Yes please.

[bergos]

Closed based on the resolution in #136