Horton

Re: Suspicious strcmp() in src/backend/parser/parse_expr.c

David Rowley <david.rowley@2ndquadrant.com>

View thread View on postgr.es
From: David Rowley <david.rowley@2ndquadrant.com>
To: Thomas Munro <thomas.munro@gmail.com>
Cc: Michael Paquier <michael@paquier.xyz>, Tom Lane <tgl@sss.pgh.pa.us>, Andrew Gierth <andrew@tao11.riddles.org.uk>, Rikard Falkeborn <rikard.falkeborn@gmail.com>, PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>
Date: 2019-04-11T03:43:49Z
Lists: pgsql-bugs 
On Thu, 11 Apr 2019 at 15:27, Thomas Munro <thomas.munro@gmail.com> wrote:
>
> On Thu, Apr 11, 2019 at 2:20 PM Michael Paquier <michael@paquier.xyz> wrote:
> > On Wed, Apr 10, 2019 at 06:43:32PM -0400, Tom Lane wrote:
> > > Indeed.  Considering how much I hate using strcmp's result as a boolean,
> > > you'd think I'd have got that right.  Thanks for noticing!
> >
> > Just a note about those strcmp() calls using a boolean as return
> > result in the tree:
> > src/backend/commands/lockcmds.c: (!strcmp(rte->eref->aliasname, "old")
>
> Don't look at contrib/spi/refint.c if you value your sanity.

hmm, yeah. Take a non-bool expression, make it into a bool expression,
then compare that to 0 to make it look like a non-bool expression.
Weird.

If we're fixing some, we may as well do them all.

-- 
 David Rowley                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

Commits

  1. Fix more strcmp() calls using boolean-like comparisons for result checks

  2. Fix backwards test in operator_precedence_warning logic.