From c98b5f094f133b5b9119c84f21351cfda0a42ff0 Mon Sep 17 00:00:00 2001 From: jnasbyupgrade Date: Mon, 10 Aug 2026 14:27:42 -0500 Subject: [PATCH] Fix object_group.sql's pgTAP plan count (24 -> 25) The "Drop tests" section runs 4 assertions (3 throws_ok, 1 lives_ok at lines 128/134/140/146) but the plan math only counted +3 for it. Traced by matching the plan formula's per-section subtotals against the actual list of 25 executed test names -- every other section's count was already correct. verify-results (make's real pgtap-aware gate, as opposed to plain pg_regress diffing) is what catches this; it was invisible before because the stale count was already baked into the committed expected-output file, so a plain diff always matched. Co-Authored-By: Claude --- test/expected/object_group.out | 3 +-- test/sql/object_group.sql | 2 +- 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/test/expected/object_group.out b/test/expected/object_group.out index d286833..c3f5eb8 100644 --- a/test/expected/object_group.out +++ b/test/expected/object_group.out @@ -1,5 +1,5 @@ \set ECHO none -1..24 +1..25 ok 1 - Register test table 1 ok 2 - object_group__create(...) for group name that is too long throws error ok 3 - object_group__create('object reference test group') @@ -25,5 +25,4 @@ ok 22 - Removing group with items in it fails ok 23 - __object__remove() for col1 works ok 24 - __object__remove() for test_table_2 works ok 25 - Removing empty group works -# Looks like you planned 24 tests but ran 25 # TRANSACTION INTENTIONALLY LEFT OPEN! diff --git a/test/sql/object_group.sql b/test/sql/object_group.sql index 6304892..a3ab46b 100644 --- a/test/sql/object_group.sql +++ b/test/sql/object_group.sql @@ -35,7 +35,7 @@ SELECT plan( +1 --TODO + 2 -- __object__add +3 -- object__getsert with group - +3 -- Drop tests + +4 -- Drop tests +4 -- __object__remove