Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Duplicate
-
12.2.2
-
None
-
Not for Release Notes
Description
The resolver path appears to assume that each outer-context transition has an
owning subquery item. This fake/top-select shape violates that invariant. The query exits with client status 137; the server log shows signal 11 with `Item_ref::fix_fields(THD*, Item**)` in the backtrace.
DROP DATABASE IF EXISTS db; |
CREATE DATABASE db; |
USE db; |
|
|
CREATE TABLE t ( |
id INT PRIMARY KEY, |
p INT, |
o INT, |
v INT |
);
|
|
|
SELECT 1 |
UNION
|
SELECT 2 |
UNION
|
SELECT ROW_NUMBER() OVER (ORDER BY t.id) |
FROM t |
JOIN t AS t_aux1 |
ON (SELECT 1 FROM dual HAVING AAA); |
Stack Trace:
mariadbd(my_print_stacktrace+0x30)[0x5a4feb2fb960]
mariadbd(handle_fatal_signal+0x1f3)[0x5a4feae47843]
/lib/x86_64-linux-gnu/libc.so.6(+0x45330)[0x716034ead330]
mariadbd(_ZN8Item_ref10fix_fieldsEP3THDPP4Item+0x1e0)[0x5a4feae860e0]
mariadbd(_ZN9Item_func10fix_fieldsEP3THDPP4Item+0x8d)[0x5a4feaeca96d]
mariadbd(+0x747e8c)[0x5a4fea9cfe8c]
mariadbd(_ZN4JOIN7prepareEP10TABLE_LISTP4ItemjP8st_orderbS5_S3_S5_P13st_select_lexP18st_select_lex_unit+0xcdc)[0x5a4feab7471c]
mariadbd(_ZN30subselect_single_select_engine7prepareEP3THD+0x691)[0x5a4feaf2fc51]
mariadbd(_ZN14Item_subselect10fix_fieldsEP3THDPP4Item+0x1af)[0x5a4feaf2a60f]
mariadbd(_Z13setup_on_exprP3THDP10TABLE_LISTb+0x95)[0x5a4feaa8b235] ...
Attachments
Issue Links
- duplicates
-
MDEV-28506 SIGSEGV's in find_field_in_table[s][_ref], Item_field::fix_fields, create_view_field and MemcmpInterceptorCommon | Assertions `(*select_ref)->fixed' or '->is_fixed' and `table_list->table' failed
-
- Stalled
-