MDEV-32401 expression cache lead to crash if table of wrong type created and the cache switched off

[forkfun]

Cherry-picked from 3d0ea3d (10.6 backport: #5492)

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ spetrunia
❌ sanja-byelkin
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[Copilot]

Copilot review overview

🟡 Changes recommended

One moderate documentation/invariant issue and two minor comment issues remain unresolved.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

• Fix typo and clarify flag column-selection behavior · New

What changed in this PR

Fixes an expression-cache crash involving aggregate result_field handling for temporary tables.

Changes:

• Controls aggregate field updates with TMP_TABLE_ALL_COLUMNS.
• Removes obsolete materialized-subquery state.
• Adds regression coverage for TEXT, CAST, and INT cases.

Review findings:

• sql/sql_priv.h: nit (3 votes)—correct the typo and wording in the new comment.
• sql/sql_priv.h: moderate (1 vote)—clarify or enforce the comment’s broader result_field invariant.
• sql/sql_select.cc: nit (1 vote)—update stale rationale describing the removed condition.

File	Reviewed change
sql/​sql_select.cc	Adjusts aggregate result-field handling.
sql/​sql_priv.h	Defines and documents the temporary-table option.
sql/​sql_class.h	Removes obsolete state.
sql/​sql_class.cc	Removes obsolete initialization.
sql/​item_subselect.cc	Removes obsolete state assignment.
mysql-test/​main/​subselect_cache.test	Adds regression queries.
mysql-test/​main/​subselect_cache.result	Records expected results.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[spetrunia]

Generally, looks like a step in the right direction.

I had considered splitting TMP_TABLE_ALL_COLUMNS into separate flags denoting each functionality but that gets too intrusive.

@sanja-byelkin , could you please check those two commits (also pushed into this PR, BB doesn't show any hits for the assertion):

commit 7519e35963612486b703a05722e6733c09918be6 (HEAD -> bb-11.4-MDEV-32401, origin/bb-11.4-MDEV-32401)
Author: Sergei Petrunia <sergey@mariadb.com>
Date:   Tue Sep 22 21:08:22 2026 +0300

    Fix comments

commit 7fa37b0760b891f0500559db99a632bb84cf3270
Author: Sergei Petrunia <sergey@mariadb.com>
Date:   Tue Sep 22 20:07:03 2026 +0300

    Add an assertion.

Any objections?