Custom Node Skipping Feature

[lread]

Have jotted some thoughts here.

Related issue is #70.

Opened this discussion to gather any feedback/ideas folks might have.

[KingMob]

I just read your doc, and I have a couple comments:

1. This is a very general solution. Have there been any requests for skipping outside of uneval/comments? If so, could those requests be adequately handled by pre/postwalk? The pareto solution that satisfies the majority of requests so far might be to handle uneval/comments, and defer more general skipping until someone needs it.
2. Searching upwards can be avoided if properties (like being unevaluated) are propagated down from parents. This might not be generally applicable, but may be a good trade-off for uneval/comment parents.

[lread]

Thanks @KingMob, I've been waiting for interested folks to chime in on this one.

The frustration I've heard over the years (and have just recently felt the awkwardness of myself) is that some users of rewrite-clj are annoyed by seeing #_ ... nodes when using the zip API. I've made the assumption that these users also might want to sometimes also skip (comment ...) nodes.

Those using rewrite-clj to parse Clojure and get at the code that would run would likely want to skip both #_ ... and (comment ...) nodes. This might be your use case.

Even if I were to expose a facility for optionally skipping also only #_ ... and/or (comment ...) nodes, I think I'd likely try to code it up generically.

The current walk support will be more appropriate for some use cases. But I think it would be nice to skip certain nodes with the zip API without using walk fns.

I think your idea of propagating down skipped-ness is a probably a good optimization.