aminghadersohi commented on PR #44148:
URL: https://github.com/apache/superset/pull/44148#issuecomment-5625317686

   Maintainer follow-up in `7975e11`: merged upstream `39944ff` normally. The 
documentation conflict retained both geographic generation and 
dataset/nested-field guidance; upstream forwarded-filter behavior and all 
geographic paths remain covered.
   
   Response to the additional suggestions in [the review 
report](https://github.com/apache/superset/pull/44148#issuecomment-5623168196):
   
   - Fixed typed world-map non-array input handling with an actionable 
translated error and four malformed-result regressions. Translated world 
metric/country errors and removed redundant numeric coercion.
   - Moved geographic ASCII rendering inside the common clamp/error boundary; 
regression cases cover all three map types at negative width and table-render 
failure.
   - Preserved temporal recommendations ahead of geographic heuristics, 
excluded boolean identifiers, and bounded identifier cardinality. Regression 
tests cover each condition; these remain recommendations, not inferred 
geographic validity.
   - Cached immutable world-country aliases by format and indexed country-map 
exact/folded aliases once per transform. Added cache/format-isolation and 
duplicate-name regressions; existing case-safe ambiguity/parity tests still 
pass.
   - Corrected Gauge rebind guidance, preserving the native Gauge wording; 
public FastMCP tests verify Gauge and all geographic rebind errors. 
Consolidated preview-type guards, compile error codes, and cross-filter/drill 
source lookup. Added the deferred-import explanation.
   - Retained strict finite metrics and frontend validation intentionally: 
saved Explore/dashboard renders can query new data without invoking MCP 
compile. Silently accepting missing/null metrics or removing client validation 
would violate the strict typed contract. Legacy maps without the typed opt-in 
retain their existing behavior.
   - Retained the bounded five-country contract rather than advertising 
unimplemented normalization. Duplicate bundled UK `Halton` names are 
deliberately rejected; documentation and Python/TypeScript regressions 
demonstrate using `GB-HAL` or `WRL` instead. We do not guess or silently 
relabel the bundled geometry.
   - The reported record-property reads do not mutate prototypes; the numeric 
checks reject nonnumeric inherited members. No prototype-pollution write or 
security-boundary violation was identified, so no speculative security fix was 
added.
   
   Validation: **4,162 MCP tests passed**, 2 skipped; **180 focused geographic 
tests passed**; **181 frontend tests passed** across 16 suites; native plugin 
TypeScript builds and full branch pre-commit/MyPy/Ruff/Pylint/type/lint checks 
passed. Exact-head CI is pending. There were no inline review threads to 
resolve; none were marked resolved.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to