Re: pgsql: Ensure that length argument of memcmp() isn't seen as negative. - Mailing list pgsql-committers

From Andres Freund
Subject Re: pgsql: Ensure that length argument of memcmp() isn't seen as negative.
Date
Msg-id 20220215224625.poqi7mrwtkhcsdpv@alap3.anarazel.de
Whole thread Raw
In response to pgsql: Ensure that length argument of memcmp() isn't seen as negative.  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-committers
Hi,

On 2022-02-15 22:28:23 +0000, Tom Lane wrote:
> Ensure that length argument of memcmp() isn't seen as negative.
> 
> I think this will shut up a weird warning from buildfarm member
> serinus.  Perhaps it'd be better to change tsCompareString's
> length arguments to unsigned, but that seems more invasive
> than is justified.

Just FYI, I'd reported this to gcc some time last year
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100442
and then forgot about it.

They're not wrong there is the possibility of exceeding the bounds, but the
error message still is quite confusing.

Greetings,

Andres Freund



pgsql-committers by date:

Previous
From: Tom Lane
Date:
Subject: Re: pgsql: Track LLVM 15 changes.
Next
From: Peter Geoghegan
Date:
Subject: pgsql: Update "don't truncate with failsafe" rationale.