[dmealing]

What Changed

• ERR_COLUMN_NAME_COLLISION skips an object whose sources are all read-only and include a source.rdb @kind: view with an @sql body. Two fields may read one column in such a view. The same exemption is applied in the TypeScript, Java, C# and Python validation passes.
• Derived views (no @sql) and objects with a writable source are still checked.
• When both colliding fields already have an explicit @column, the error message now says to rename one field or point one @column at a different column. The error code is unchanged.
• Adds conformance fixtures column-name-shared-in-sql-view-ok, error-column-name-collision-derived-view and error-column-name-collision-table-with-sql-view, plus a per-port regression test (TypeScript, Java, C#, Python). ERROR-CODES.json and docs/CONFORMANCE.md are updated to match, and the CHANGELOG Fixed entry is added.

Validation

Risk Assessment

✅ Low: Narrow, symmetric exemption in all four ports (every source read-only and at least one @sql view), backed by shared conformance fixtures and per-port message tests. It matches the intent and the recorded decisions, and no wrong-result path turned up.

Testing

The Test agent exceeded its invocation budget before live validation completed; no evidence was gathered for this head.

• Outcome: ⚠️ 1 warning across 1 run (1h56m8s)

Pipeline

Updates from git push no-mistakes

✅ **Intent** - passed

✅ No issues found.

✅ **Rebase** - passed

✅ No issues found.

✅ **Review** - passed

✅ No issues found.

⚠️ **Test** - 1 warning

• ⚠️ The Test agent did not finish within its invocation budget. Reported: agent run tests timed out at its 1h30m0s still-working cap (silent budget 1h15m0s; ran 1h30m2s): agent last produced output 23m20s ago (3 observed); agent reported: claude exited: signal: terminated: glm-claude: step=test route=light glm-claude: trying haiku55 (effort low, thinking off)... glm-claude: haiku55 failed (timed out after 2000s, partial stream discarded) glm-claude: haiku55 failed — failing over glm-claude: trying sonnet55 (effort high)... glm-claude: sonnet55 failed (timed out after 2000s, partial stream discarded) glm-claude: sonnet55 failed — fa.... This is a budget or provider-slowness cut, not a code failure. Re-running the same request costs another full budget, so no further attempt is made automatically. If this repository's targeted tests or evidence gathering routinely stay quiet longer than the default 30m0s, raise test_agent_timeout in global config. If the turn was still producing output or running a child process, set or raise test_agent_working_timeout there instead: a turn runs past test_agent_timeout only up to that still-working cap. Respond with fix to spend another budget: a repair turn runs only for selected findings other than this budget cut, then validation re-runs. Or abort and retry after raising the budget.

✅ **Document** - passed

✅ No issues found.

✅ **Lint** - passed

✅ No issues found.

✅ **Push** - passed

✅ No issues found.