Re: Cast to uint16 in pg_checksum_page()

On 3/4/20 1:05 AM, Tom Lane wrote:
> Michael Paquier <michael(at)paquier(dot)xyz> writes:
>> On Tue, Mar 03, 2020 at 06:37:36PM -0500, David Steele wrote:
>>> It seems like an explicit cast to uint16 would be better?
>
>> Attempting to compile the backend code with -Wconversion leads to many
>> warnings, still there has been at least one fix in the past to ease
>> the use of the headers in this case, with b5b3229 (this made the code
>> more readable). Should we really care about this case?
>
> Per the commit message for b5b3229, it might be worth getting rid of
> such messages for code that's exposed in header files, even if removing
> all of those warnings would be too much work. Perhaps David's use-case
> is an extension that's using that header?

Yes, this is being included in an external project. Previously we have
used a highly marked-up version but we are now trying to pull in the
header more or less verbatim.

Since this header is specifically designated as something external
projects may want to use I think it makes sense to fix the warning.

Regards,
--
-David
david(at)pgmasters(dot)net