Bug #115498 Test main.opt_hints_hypergraph always failing
Submitted: 3 Jul 2024 11:09 Modified: 3 Jul 2024 13:07
Reporter: Laurynas Biveinis (OCA) Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Server: Tests Severity:S7 (Test Cases)
Version:9.0 OS:MacOS
Assigned to: CPU Architecture:ARM

[3 Jul 2024 11:09] Laurynas Biveinis
Description:
[ 50%] main.opt_hints_hypergraph                 [ fail ]
        Test ended at 2024-07-03 14:08:16

CURRENT_TEST: main.opt_hints_hypergraph
--- /Users/laurynas/vilniusdb/mysql-9.0.0/mysql-test/r/opt_hints_hypergraph.result	2024-07-01 14:55:47
+++ /Users/laurynas/vilniusdb/mysql-9.0.0/_build-debug/mysql-test/var/log/opt_hints_hypergraph.reject	2024-07-03 14:08:15
@@ -1257,21 +1257,21 @@
 -> Inner hash join (alias10.f1 = alias11.f1)  (rows=1)
     -> Table scan on alias10  (rows=1)
     -> Hash
-        -> Inner hash join (cast(alias4.f2 as double) = cast(alias11.f2 as double))  (rows=1)
-            -> Left hash join (alias4.f2 = alias5.f2)  (rows=1)
+        -> Left hash join (alias4.f2 = alias5.f2)  (rows=1)
+            -> Inner hash join (cast(alias4.f2 as double) = cast(alias11.f2 as double))  (rows=1)
                 -> Table scan on alias4  (rows=1)
                 -> Hash
-                    -> Left hash join (alias5.f1 = alias8.f2)  (rows=1)
-                        -> Inner hash join (alias5.f2 = alias6.f1)  (rows=1)
-                            -> Table scan on alias5  (rows=1)
-                            -> Hash
-                                -> Table scan on alias6  (rows=1)
+                    -> Table scan on alias11  (rows=1)
+            -> Hash
+                -> Left hash join (alias5.f1 = alias8.f2)  (rows=1)
+                    -> Inner hash join (alias5.f2 = alias6.f1)  (rows=1)
+                        -> Table scan on alias5  (rows=1)
                         -> Hash
-                            -> Inner hash join (<hash>(alias7.f1)=<hash>(alias8.f1)), extra conditions: (alias7.f1 = alias8.f1)  (rows=1)
-                                -> Table scan on alias7  (rows=1)
-                                -> Hash
-                                    -> Table scan on alias8  (rows=1)
-            -> Hash
-                -> Table scan on alias11  (rows=1)
+                            -> Table scan on alias6  (rows=1)
+                    -> Hash
+                        -> Inner hash join (<hash>(alias7.f1)=<hash>(alias8.f1)), extra conditions: (alias7.f1 = alias8.f1)  (rows=1)
+                            -> Table scan on alias7  (rows=1)
+                            -> Hash
+                                -> Table scan on alias8  (rows=1)
 
 DROP TABLES t2, t4, t5, t6, t7, t10, t11;

mysqltest: Result content mismatch

How to repeat:
CMake options -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DCMAKE_BUILD_TYPE=Debug -DWITH_DEBUG=ON -DMYSQL_MAINTAINER_MODE=ON -DWITH_SYSTEM_LIBS=ON -DWITH_NDBCLUSTER_STORAGE_ENGINE=OFF -DFORCE_COLORED_OUTPUT=ON -DWITH_ZLIB=bundled -DCMAKE_CXX_FLAGS=-Wno-unused-lambda-capture -DCMAKE_C_FLAGS_DEBUG=-g -Wno-unused-lambda-capture -DCMAKE_CXX_FLAGS_DEBUG=-g -Wno-unused-lambda-capture -DCMAKE_CXX_FLAGS_RELEASE=-O2 -g -DNDEBUG -Wno-unused-lambda-capture
[3 Jul 2024 12:11] MySQL Verification Team
Hi Mr. Biveinis,

Would you be so kind to share with us your findings regarding the options that you use.

Which option is the one that causes the error, beside the options for Debug builds ......

Thanks in advance ......
[3 Jul 2024 13:07] MySQL Verification Team
Hi Mr. Biveinis,

Thank you for your bug report.

We managed to reproduce the problem, but only partially ..........

However, this is enough to verify this report.

We do not know whether yoru compilation flags are to blame , but we shall leave it on Development to analyse this.

Verified as reported.
[23 Nov 2024 2:26] Jon Stephens
Fixed in MySQL 9.2.0 by BUG#116391.

Closed.
[25 Nov 2024 10:36] MySQL Verification Team
Thank you, Jon.