mirror of https://github.com/citusdata/citus.git
PG18 - Normalize PG18 EXPLAIN: hide “Storage … Maximum Storage …” line (#8292)
fixes #8267 * Extend `src/test/regress/bin/normalize.sed` to drop the new PG18 EXPLAIN instrumentation line: ``` Storage: <Memory|Disk|Memory and Disk> Maximum Storage: <size> ``` which appears under `Materialize`, some `CTE Scan`s, etc. when `ANALYZE` is on. **Why** * PG18 added storage usage reporting for materialization/tuplestore nodes. It’s useful for humans but creates noisy, non-semantic diffs in regression output. There’s no EXPLAIN flag to suppress it, so we normalize in tests instead. This PR wires that normalization into our sed pipeline. https://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=1eff8279d **How** * Add a narrowly scoped sed rule that matches only lines starting with `Storage:` (keeping `Sort Method`, `Hash`, `Buffers`, etc. intact). Use ERE compatible with `sed -Ef` and Python `re` (no POSIX character classes), e.g.: ``` /^[ \t]*Storage:[ \t].*$/d ```pull/8307/head
parent
61b491f0f4
commit
be2fcda071
|
|
@ -378,3 +378,7 @@ s/\<is referenced from table\>/is still referenced from table/g
|
||||||
|
|
||||||
# PG18: EXPLAIN ANALYZE prints "Index Searches: N" for index scans — remove it
|
# PG18: EXPLAIN ANALYZE prints "Index Searches: N" for index scans — remove it
|
||||||
/^\s*Index Searches:\s*\d+\s*$/d
|
/^\s*Index Searches:\s*\d+\s*$/d
|
||||||
|
|
||||||
|
# EXPLAIN (PG18+): hide Materialize storage instrumentation
|
||||||
|
# this rule can be removed when PG18 is the minimum supported version
|
||||||
|
/^[ \t]*Storage:[ \t].*$/d
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue