Fix fill-outline-color fallback for feature-state expressions

[itisyb]

Summary

• preserve fill-outline-color fallback to fill-color when feature-state is missing so fill rendering does not break during initial paint population
• apply the same fallback behavior to both source and composite expression binders, and refresh it after paint updates so later fill-color changes stay in sync
• add unit coverage for missing state, feature-state updates, fallback refreshes, and invalid color warnings

Testing

• npm run lint -- src/data/program_configuration.ts src/data/program_configuration.test.ts
• npm run test-unit -- src/data/program_configuration.test.ts

Notes

• I attempted to add a render regression too, but npm run build-dev is currently blocked by unrelated existing TypeScript errors in src/data/bucket/line_bucket.ts and src/source/geojson_*, so this PR keeps the regression coverage focused on unit tests.

Fixes #4234

[itisyb]

Also created a small demo but don't think need to push/show for this case

[codecov]

Codecov Report

❌ Patch coverage is 90.56604% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.58%. Comparing base (789b3da) to head (e230a83).
⚠️ Report is 23 commits behind head on main.

Files with missing lines	Patch %	Lines
src/data/program_configuration.ts	90.56%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7420      +/-   ##
==========================================
- Coverage   92.86%   92.58%   -0.29%     
==========================================
  Files         288      289       +1     
  Lines       23998    24060      +62     
  Branches     5095     5118      +23     
==========================================
- Hits        22286    22276      -10     
- Misses       1712     1784      +72     

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

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[HarelM]

Thanks for taking the time to open this PR.
I don't think there's an issue with building or render test as CI passes without any issues as far as I know.
From the initial post it's hard to understand what the changes are and why these are the right places to do them, can you elaborate a bit more on the solution?

[itisyb]

Thanks for taking the time to open this PR. I don't think there's an issue with building or render test as CI passes without any issues as far as I know. From the initial post it's hard to understand what the charges are and why these are the right places to do them, can you elaborate a bit more on the solution?

Thanks for taking the time to open this PR. I don't think there's an issue with building or render test as CI passes without any issues as far as I know. From the initial post it's hard to understand what the charges are and why these are the right places to do them, can you elaborate a bit more on the solution?

Before the fix, there was a problem with how fill-outline-color was handled when using feature-state.

If the feature-state value was missing, the expression could return null. The rendering system didn’t treat that as “no value” - it treated it like an actual color. Then when packColor() tried to use it, it attempted to access color.r on null, which caused a runtime error: “Cannot read properties of null (reading 'r')”.

it already had a fallback rule where fill-outline-color defaults to fill-color if not specified. But that logic only applied in the constant (non–data-driven) case. In the feature-state (data-driven) path, that fallback wasn’t applied, so instead of defaulting correctly, it crashed.

[HarelM]

This looks too specific to fill color, are you sure this is the right place to add this code? I would expect to see this code in the fill bucket/fill program or something related to fill, and not in this "shared" program configuration code.

[HarelM]

Seems like this is still relevant after the recent force-push...

[HarelM]

Is there a way to avoid this as any?