Inaccurate comment, for family-1 polymorphic UNKNOWN type arguments
Hi Hackers,
I observed that we have inaccurate comment in
enforce_generic_type_consistency.
if (!OidIsValid(elem_typeid))
{
if (allow_poly)
{
elem_typeid = ANYELEMENTOID;
array_typeid = ANYARRAYOID;
range_typeid = ANYRANGEOID;
}
else
{
/*
* Only way to get here is if all the
polymorphic args have
* UNKNOWN inputs
*/
*ereport(ERROR, *
...............
}
........
}
We reach the error condition even if there is any "anycompatible" parameter
is present (and that is of some known type say int32).
I think developer intend to report error if "we have all the family-1
polymorphic arguments as UNKNOWN".
Thoughts?
Please find attached the patch to fix this typo.
Thanks,
Himanshu
EnterpriseDB: http://www.enterprisedb.com
Attachments:
v1-0001-FIX-for-incorrect-comment-for-UNKNOWN-input-type-.patchtext/x-patch; charset=US-ASCII; name=v1-0001-FIX-for-incorrect-comment-for-UNKNOWN-input-type-.patchDownload+1-2
Himanshu Upadhyaya <upadhyaya.himanshu@gmail.com> writes:
I observed that we have inaccurate comment in
enforce_generic_type_consistency.
Fair point, but that's not the only old comment that's not being careful
about it. I applied the attached.
regards, tom lane