sql: fix labelled tuple type checking

[mgartner]

Fixes #106303

Release note (bug fix): A bug has been fixed that caused internal errors
instead of user errors when queries contained labelled tuples with a
different number of elements and labels, e.g., (ROW(1, 2) AS a). This
bug was present since v23.1.0.

[cockroach-teamcity]

This change is 

[DrewKimball]

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @mgartner, @msirek, and @qiyanghe1998)

---

pkg/sql/sem/tree/type_check.go line 2802 at r1 (raw file):

	// mimics the check in (*Tuple).TypeCheck.
	firstLen := len(first.Exprs)
	if len(first.Labels) > 0 && len(first.Labels) != firstLen {

Should we check other tuples than the first? What happens for a test like:

SELECT CASE WHEN False THEN ROW() ELSE (ROW() AS a) END;

[msirek]

Reviewed all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @mgartner and @qiyanghe1998)

[mgartner]

Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @DrewKimball and @qiyanghe1998)

---

pkg/sql/sem/tree/type_check.go line 2802 at r1 (raw file):

Previously, DrewKimball (Drew Kimball) wrote…

Should we check other tuples than the first? What happens for a test like:

SELECT CASE WHEN False THEN ROW() ELSE (ROW() AS a) END;

Good call. Done.

[DrewKimball]

Reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (and 1 stale) (waiting on @qiyanghe1998)

[qiyanghe1998]

Reviewable status: complete! 2 of 0 LGTMs obtained (and 1 stale) (waiting on @DrewKimball)

[mgartner]

TFTRs!

bors r+

[craig]

Build succeeded:

• Bazel Essential CI (Cockroach)