fix: emission of events from root validators and extraneous emission of dependent fields

[tlambert03]

fixes #233

@andy-sweet, this implements option number 2 from #233 (comment)

• property_dependencies can include any field names ... not just property setters
• values of dependents are checked before and after setting and events are only emitted if necessary

[codspeed-hq]

CodSpeed Performance Report

Merging #234 will not alter performance

_{Comparing tlambert03:fix-root-validators (c85743f) with main (1255291)}

Summary

✅ 66 untouched benchmarks

[codecov]

Codecov Report

Patch coverage: 100.00% and no project coverage change.

Comparison is base (1255291) 100.00% compared to head (c85743f) 100.00%.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #234   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           22        22           
  Lines         1829      1843   +14     
=========================================
+ Hits          1829      1843   +14     

Files Changed	Coverage Δ
src/psygnal/_evented_model_v1.py	100.00% <100.00%> (ø)
src/psygnal/_evented_model_v2.py	100.00% <100.00%> (ø)

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[andy-sweet]

Looks good to me. Thanks for the fast fix. Just made a couple of nit-picking comments on implementation comment naming, which are optional.

[andy-sweet]

As discussed in #233, it probably also makes sense to rename property_dependencies with this solution. That doesn't have to happen in this PR, but there should be an issue to track it if this is merged without.

[tlambert03]

agreed, did you have a thought on the exact name? As mentioned over there:

I like dependencies, but since the namespace is shared with all the other pydantic stuff, perhaps event_dependencies is safer/clearer? or do you like field_dependencies better than event_dependencies?