Closed ericprud closed 2 years ago
Are there new tests that show the advantage of this indirection rather than just the extra overhead of indirection?
I don't think the tests show it off. The ShExJ.jsg is a ton simpler.
I think the strongest case comes from diving into an implementation and seeing the number of potential bugs that disappear. There are lots of places where I wasn't conditionally traversing the ShapeDecls's .shexExpr
(e.g.) because I hadn't happened to hit those lines with a schema that needed a ShapeDecl (i.e. had .abstract
set to true).
Tx @gkellogg . I'll wait 'till Iovka gets shex-java running with the current extensd tests before pushing this upstream with the idea of making her steps to passing the tests more incremental. @iovka, that may require you to do extra coding so if you prefer, just you can merge this into extends and and pull the merged extends.
Labra vocalized his approval in EXTENDS meeting on 2022-07-05
per shexSpec/spec#50