linhongyu510 commented on PR #2559: URL: https://github.com/apache/age/pull/2559#issuecomment-6144280284
Verified the default-label edge fix end-to-end in a clean Docker environment (real `pg_regress`, no AST/static shortcuts). **Environment** - Image: `postgres:18` (Debian), `postgresql-server-dev-18`, bison 3.8.2, flex 2.6.4 - PostgreSQL **18.6**; AGE built via `make PG_CONFIG=/usr/bin/pg_config install`; regression via the project harness `pg_regress --load-extension=age --temp-instance ...` (the `expr` suite). **Baseline (pre-fix `4d5707d` + this PR's `expr.sql`/`expr.out`) — RED** The new `MATCH ()-[r:_ag_label_edge]->() WHERE type(r)='_ag_label_edge'` case is elided and returns a child edge label instead of no rows. `regression.diffs`: ``` -(0 rows) + "KNOWS" +(1 row) ``` (1 of 42 tests failed: `expr`.) **Head `3d3741b3` — GREEN** With the `IS_DEFAULT_LABEL_EDGE(static_label)` guard, that case now returns the expected `(0 rows)` and `expr` passes. The only production-code delta between baseline and head is the guard in `is_redundant_edge_type_qual`; the regression SQL/expected is exactly the case requested in review. **Note on a separate flaky test** On this arm64 / PG 18.6 container, `cypher_merge` (issue_1446 label counts) prints group rows in a different order (`hub/shared` before `X/Y/Z`) with no `ORDER BY`. That query has no `type()` filter and is untouched by this change; it's pre-existing hash-aggregation ordering nondeterminism, not a regression from this PR. -- 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]
