feat(tree): validate YANG unique list constraints

[steiler]

Summary

• Add validation_entry_unique.go: walks list children, collects leaf values per unique statement, reports violations
• Wire into validation_dispatch.go and ValidationStats
• Add test YANG schema (sdcio_model_list_unique.yang) covering single-leaf, multi-leaf composite, and descendant-path unique statements
• Regenerate tests/sdcioygot/sdcio_schema.go with updated schema

Closes #445

Related

• sdc-protos PR: feat(schema): add UniqueConstraint message and unique_constraints to ContainerSchema sdc-protos#121 (adds UniqueConstraint message + unique_constraints field)
• schema-server PR: feat(schema): populate unique constraints from YANG list entries schema-server#242 (populates unique_constraints from YANG unique statements)

[codecov]

Codecov Report

❌ Patch coverage is 79.20792% with 21 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
pkg/tree/ops/validation/validation_entry_unique.go	82.10%	10 Missing and 7 partials ⚠️
pkg/config/validation.go	0.00%	2 Missing ⚠️
pkg/tree/types/validation_stats.go	0.00%	2 Missing ⚠️

📢 Thoughts on this report? Let us know!

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[alexandernorth]

should these not be in one unique entry?

[steiler]

According to the YANG 1.1 Specification (RFC 7950, Section 7.8.3), the unique statement can be used multiple times within a single list definition.

Because each unique statement enforces a distinct constraint on a different set of leafes, our validation logic must iterate through all instances of the statement rather than assuming a single occurrence per list.

[alexandernorth]

so there is no difference between

unique "ip port xyz"

and

unique "ip port"
unique "xyz"

?

I couldn't get this from the RFC

[steiler]

so there is no difference between

unique "ip port xyz"

and

unique "ip port"
unique "xyz"

?

I couldn't get this from the RFC

No — they are not the same. Each unique statement is one constraint; multiple unique statements mean multiple independent constraints.

unique "ip port xyz" (one constraint)

This enforces uniqueness of the composite (ip, port, xyz) across list entries (with the usual RFC rules for missing/empty values). Two entries may share the same ip and port if xyz differs, because it is the full triple that must not repeat.

unique "ip port" plus unique "xyz" (two constraints)

1. unique "ip port" — the pair (ip, port) must not repeat. Same ip+port on two rows is invalid even if xyz differs.
2. unique "xyz" — xyz must not repeat by itself, even if ip/port differ.

So the second form is not equivalent: it adds a separate uniqueness requirement on (ip, port) and another on xyz alone.

Tiny example

Assume all three leaves are set:

Scenario	Row A	Row B	unique "ip port xyz"	unique "ip port" + unique "xyz"
Same ip,port; different xyz	1,1,a	1,1,b	OK	Fails unique "ip port"
Different ip,port; same xyz	1,1,x	2,2,x	OK	Fails unique "xyz"

RFC reference: RFC 7950 §7.8.3 (unique statement).