Non-flowing :and parser

[frenchy64]

Close #1166

Could also provide a separate schema (called and- in spec2).

Approach from spec2

Original slack thread

TODO

• unparsing (why is this invalid?)

               (let [s [:and {:parse :non-flowing}
                        [:catn ["a" :int] ["b" :keyword]]
                        [:fn vector?]]]
                 (m/unparse s (m/parse s [3 :x])))

[opqdonut]

I'm not sure I see who would ever want the current behaviour of :and + parse? It seems to me it would be pretty difficult to use the current :and so that it works both when validating and parsing. I guess most users only validate, and making :and non-flowing wouldn't affect them?

[frenchy64]

Yes, the only case I can think of is [:and [:fn vector?] [:catn ...]] where the simpler parsing goes first and flows to the more complex one.

[frenchy64]

After some thought, I agree it makes a lot of sense to only parse the first conjunct. This also aligns with the :and generator, where the first conjunct provides the generator and the rest filter via their validators. I'll sketch it out in another PR.

[frenchy64]

Superseded by #1182