Bernd Helmle <[EMAIL PROTECTED]> writes:
> --On Mittwoch, Juni 08, 2005 14:49:56 -0400 Tom Lane <[EMAIL PROTECTED]>
> wrote:
>> Applying "const" to pointers that point to things that are not const,
>> as in
>>
>> + void
>> + ApplyTypeNamespace( Oid typeOid,
>> + const Relation rel,
>>
>> seems to me to be horrible style, even if the compiler lets you do it.
>> It's too easy to misread it as a promise not to alter the pointed-to
>> object.
> Well, i thought there *should* be a promise, not to alter *rel in that
> specific case.
Hmm? You're planning to write into the relation in question. It's
hardly likely that the structure can be expected to remain virgin...
in practice I don't think we guarantee that even for read operations.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
