Feature request - extend 'match each' to work with 'contains deep'

[anupamck]

Currently, match each doesn't work with contains deep.

Scenario:
    * def original = [ { a: 1, arr: [ { b: 2, c: 3 }, { b: 3, c: 4 } ] } ]
    * def expected = { a: 1, arr: [ { b: '#number', c: '#number' } ] }
    # This fails
    * match each original contains deep expected

This is in line with the documentation, given that there is a section for match contains deep, but no corresponding section for match each contains deep. However, I think this feature would be of great use, especially in performing schema validation for complex JSON arrays with the powerful flexibility of contains deep.

[ptrthomas]

tagging as help wanted - should be easy if you refer fix for #2093

[vincent-hennig]

I did a bit of digging together with @anupamck and it seems to us that most of the functionality for each contains deep is already implemented. We found that MatchStep.getType seems to simply be missing a case for this.

karate/karate-core/src/main/java/com/intuit/karate/MatchStep.java

Lines 133 to 162 in 694c92a

	private static Match.Type getType(boolean each, boolean not, boolean contains, boolean only, boolean any, boolean deep) {
	if (each) {
	if (contains) {
	if (only) {
	return Match.Type.EACH_CONTAINS_ONLY;
	}
	if (any) {
	return Match.Type.EACH_CONTAINS_ANY;
	}
	return not ? Match.Type.EACH_NOT_CONTAINS : Match.Type.EACH_CONTAINS;
	}
	return not ? Match.Type.EACH_NOT_EQUALS : Match.Type.EACH_EQUALS;
	}
	if (contains) {
	if (only) {
	return deep ? Match.Type.CONTAINS_ONLY_DEEP : Match.Type.CONTAINS_ONLY;
	}
	if (any) {
	return Match.Type.CONTAINS_ANY;
	}
	if (deep) {
	if (not) {
	throw new RuntimeException("'!contains deep' is not yet supported, use 'contains deep' instead");
	}
	return Match.Type.CONTAINS_DEEP;
	}
	return not ? Match.Type.NOT_CONTAINS : Match.Type.CONTAINS;
	}
	return not ? Match.Type.NOT_EQUALS : Match.Type.EQUALS;
	}

Here is the Change that I would propose:

vincent-hennig@24564d2#diff-b849d563bd2a8fccf84bcb162901dc2ba61ab6aeaebc743e0648361f2db019a4

[ptrthomas]

@vincent-hennig sounds right to me ! how about you attempt a PR ? if all existing tests pass, we are all good. and you can add a couple of tests for this case

[ptrthomas]

1.3.1 released