Fixed PL/pgSQL panic and part of issue #2197

[Hydrocharged]

This fixes some bugs found from the import dumps, alongside a portion of:

• Bugs with table composite types #2197

The remainder of the above issue will be covered in another PR

[github-actions]

	Main	PR
covering_index_scan_postgres	571.15/s	562.62/s	-1.5%
index_join_postgres	124.76/s	123.54/s	-1.0%
index_join_scan_postgres	187.11/s	186.18/s	-0.5%
index_scan_postgres	11.49/s	11.41/s	-0.7%
oltp_point_select	2345.42/s	2370.12/s	+1.0%
oltp_read_only	1722.50/s	1712.91/s	-0.6%
select_random_points	111.86/s	111.89/s	0.0%
select_random_ranges	424.71/s	431.31/s	+1.5%
table_scan_postgres	11.15/s	11.27/s	+1.0%
types_table_scan_postgres	5.21/s	5.24/s	+0.5%

[github-actions]

	Main	PR
Total	42090	42090
Successful	17565	17568
Failures	24525	24522
Partial Successes1	5583	5583

	Main	PR
Successful	41.7320%	41.7391%
Failures	58.2680%	58.2609%

${\color{lightgreen}Progressions (2)}$

arrays

QUERY: select unnest(array[1,2,3,4.5]::float8[]);

QUERY: select unnest(array[1,2,3,4.5]::numeric[]);

Footnotes

1. These are tests that we're marking as Successful, however they do not match the expected output in some way. This is due to small differences, such as different wording on the error messages, or the column names being incorrect while the data itself is correct. ↩

[fulghum]

Looks good!

[Hydrocharged]

@fulghum has rebased his PR (#2198) onto this one, so I'll close this since the changes will be merged with his.