ui: route partial-response localStorage through a shared accessor

[donaldraph]

• I added CHANGELOG entry for this change.
• Change is not relevant to the end user.

Changes

Every localStorage read/write in the React app goes through the useLocalStorage hook, except the usePartialResponse key in Panel.tsx and PanelList.tsx, which touch localStorage directly. Panel is a class component and can't call the hook, which is why it bypassed it.

I pulled the raw operations out into three plain functions in useLocalStorage.tsx — getStorageItem, setStorageItem, hasStorageItem — so both hook and non-hook callers go through the same accessor. useLocalStorage becomes a thin wrapper over them, and I routed the three bypass sites through them instead of touching localStorage directly. After this, no raw localStorage. access remains in the app outside that one module.

I intended this as a read/write refactor with no behavior change, so I checked the full git history of the key: it has only ever been written as JSON.stringify(<boolean>) since it was introduced, so under normal use the stored value is always "true", "false", or absent.

One thing I want to flag: routing reads through JSON.parse (which the old raw-string comparison didn't do) would throw on a non-JSON value. I confirmed that can't happen through app code, only via manual localStorage tampering or corruption — but to keep the old "never crashes on a bad value" behavior, I added a try/catch in getStorageItem that falls back to the initial value on a parse error, the same approach jaeger-ui uses for this. I used hasStorageItem (no parsing) for the componentDidMount existence check to keep exact parity there.

One residual I'll disclose: handleChangePartialResponse now parses via the typed accessor, so a whitespace-padded value like " true " would parse to true where the old === 'true' didn't match. No app code has ever produced such a value, and closing it fully would mean abandoning the typed accessor for that one read, which didn't seem worth it — happy to change that if you'd prefer.

Verification

I verified behavior parity empirically across every value the key could hold before settling on this shape. make react-app-lint and make react-app-test both pass (260 tests).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 64.52%. Comparing base (51afa7e) to head (6a8c1e4).
⚠️ Report is 21 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8992      +/-   ##
==========================================
- Coverage   64.53%   64.52%   -0.02%     
==========================================
  Files         289      289              
  Lines       37366    37366              
==========================================
- Hits        24113    24109       -4     
- Misses      11156    11161       +5     
+ Partials     2097     2096       -1     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.