Unify selector syntax - Githubissues

ietf-wg-jsonpath / draft-ietf-jsonpath-base

Development of a JSONPath internet draft

https://ietf-wg-jsonpath.github.io/draft-ietf-jsonpath-base/

Other

59 stars 20 forks source link

Unify selector syntax #259

Closed glyn closed 2 years ago

glyn commented 2 years ago

This draft PR attempts to extract the syntactic advantages from PR https://github.com/ietf-wg-jsonpath/draft-ietf-jsonpath-base/pull/258 while preserving the semantic structure. It's only a start, but it gives an idea of a purely syntactic approach to issue #201. The semantic structure and document structure are unchanged.

To do:

[ ] Shorthand syntax as in PR https://github.com/ietf-wg-jsonpath/draft-ietf-jsonpath-base/pull/258

Fixes https://github.com/ietf-wg-jsonpath/draft-ietf-jsonpath-base/issues/201

gregsdennis commented 2 years ago

while minimising the semantic disruption

There should be no semantic disruption in #258 as it is a non-functional change.

glyn commented 2 years ago

while minimising the semantic disruption

There should be no semantic disruption in #258 as is a non-functional change.

I agree that #258 aims to preserve the current JSONPath semantics. The disruption is, at least from my perspective, in the way the semantics is defined in #258 with the splits between child selector and various *pickers.

goessner commented 2 years ago

We agreed, that Greg should elaborate his approach of strict distinction between *pickers and selectors, which he did with #258.

Now we need to discuss the result, which might be the main topic of the next interim meeting. Some aspects:

I doubt, that we can say: "One approach is more correct than the other". They are only (fundamentally) different.
We urgently need to term that *pickers beast. I already made a proposal here.
I would prefer to say 'pickers match nodes' over 'pickers match values'.
I am not so happy with calling *pickers 'matchers'.
Glyn made two proposal to define, how *pickers and selectors relate and interoperate. They go into the right direction and to get them straight and make one out of them is quite important.

-- Stefan

glyn commented 2 years ago

The purpose of this draft PR was to test a possible consensus around a purely syntactic unification of child selectors. It seems that most people (except me) would like to continue with the approach of PR #258 and define *picker semantics. I'll therefore close this PR and look to PR #258 to gain (rough - possibly excluding me) consensus.