Horton

Re: BUG #17858: ExecEvalArrayExpr() leaves uninitialised memory for multidim array with nulls

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alexander Lakhin <exclusion@gmail.com>
Cc: Richard Guo <guofenglinux@gmail.com>, pgsql-bugs@lists.postgresql.org
Date: 2023-03-26T17:44:41Z
Lists: pgsql-bugs 
Alexander Lakhin <exclusion@gmail.com> writes:
> So maybe the fix for the bug can be seen as a supplement for 18c0b4ecc.

Yeah, I agree.  Given that precedent, there's little excuse for
ExecEvalArrayExpr not to use palloc0; the fact that it doesn't
already was just an oversight in 18c0b4ecc.

I realized that there's a completely separate way in which
ExecEvalArrayExpr is a few bricks shy of a load: it fails to check
for overflow of the output array size.

Pushed the discussed changes along with a fix for that.

			regards, tom lane

Commits

  1. Fix oversights in array manipulation.

  2. Fix array- and path-creating functions to ensure padding bytes are zeroes.