Horton

Re: Making empty Bitmapsets always be NULL

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Richard Guo <guofenglinux@gmail.com>
Cc: Nathan Bossart <nathandbossart@gmail.com>, pgsql-hackers@lists.postgresql.org
Date: 2023-03-02T17:04:27Z
Lists: pgsql-hackers 
Richard Guo <guofenglinux@gmail.com> writes:
> It seems that the Bitmapset checked by bms_is_empty_internal cannot be
> NULL from how it is computed by a function.  So I wonder if we can
> remove the check of 'a' being NULL in that function, or reduce it to an
> Assert.

Yeah, I think just removing it is sufficient.  The subsequent attempts
to dereference the pointer will crash just fine if it's NULL; we don't
need an Assert to help things along.

> It seems that in create_lateral_join_info around line 689, the
> bms_is_empty check of lateral_relids is not necessary, since we've
> checked that lateral_relids cannot be NULL several lines earlier.

Good catch, I missed that one.

Pushed, thanks for reviewing.

			regards, tom lane

Commits

  1. Remove trailing zero words from Bitmapsets

  2. Remove local optimizations of empty Bitmapsets into null pointers.

  3. Require empty Bitmapsets to be represented as NULL.

  4. Mop up some undue familiarity with the innards of Bitmapsets.

  5. Remove bms_first_member().