Horton

Thread

Commits

  1. Fix accidental use of = instead of ==

  2. Change gist stratnum function to use CompareType

  1. Accidental assignment instead of compare in GetOperatorFromCompareType?

    Jacob Brazeal <jacob.brazeal@gmail.com> — 2025-03-03T05:05:47Z

    Hi all,

It looks like commit 630f9a43cece93cb4a5c243b30e34abce6a89514 omitted an
equals sign in an ereport() in GetOperatorFromCompareType, giving the line:

cmptype = COMPARE_EQ ? errmsg("could not identify an equality operator for
type %s", format_type_be(opcintype)) :

I think the impact is just that we'd never see the error messages for
the COMPARE_OVERLAP and COMPARE_CONTAINED_BY cases logged.

Regards,
Jacob

  2. Re: Accidental assignment instead of compare in GetOperatorFromCompareType?

    Jacob Brazeal <jacob.brazeal@gmail.com> — 2025-03-03T05:06:18Z

    Sorry, I attached the wrong patch.

>

  3. Re: Accidental assignment instead of compare in GetOperatorFromCompareType?

    Michael Paquier <michael@paquier.xyz> — 2025-03-03T05:26:44Z

    On Sun, Mar 02, 2025 at 09:06:18PM -0800, Jacob Brazeal wrote:
> Sorry, I attached the wrong patch.
                     errcode(ERRCODE_UNDEFINED_OBJECT),
-                    cmptype = COMPARE_EQ ? errmsg("could not identify an equality operator for type %s", format_type_be(opcintype)) :
+                    cmptype == COMPARE_EQ ? errmsg("could not identify an equality operator for type %s", format_type_be(opcintype)) :

Yep, you're right.  That's a typo coming from the recent commit
630f9a43cece.
--
Michael

  4. Re: Accidental assignment instead of compare in GetOperatorFromCompareType?

    Peter Eisentraut <peter@eisentraut.org> — 2025-03-04T08:50:36Z

    On 03.03.25 06:05, Jacob Brazeal wrote:
> Hi all,
> 
> It looks like commit 630f9a43cece93cb4a5c243b30e34abce6a89514 omitted an 
> equals sign in an ereport() in GetOperatorFromCompareType, giving the line:
> 
> cmptype = COMPARE_EQ ? errmsg("could not identify an equality operator 
> for type %s", format_type_be(opcintype)) :
> 
> I think the impact is just that we'd never see the error messages for 
> the COMPARE_OVERLAP and COMPARE_CONTAINED_BY cases logged.

Fixed, thanks.