On 09/22/2013 02:17 PM, Fabien COELHO wrote:
>
> Hello,
>
>> There is no pg_sleep(text) function and the cast is unknown->double
>> precision.
>
> My mistake.
>
> As I understand it, pg_sleep('12') currently works and would not
> anymore once your patch is applied. That is the concern raised by
> Robert Haas.
That is correct.
>
>>> ISTM that providing "pg_sleep(TEXT)" cleanly resolves the
>>> upward-compatibility issue raised.
>>
>> I don't like this idea at all. If we don't want to break compatibility
>> for callers that quote the value, I would rather the patch be rejected
>> outright.
>
> That was just a suggestion, and I was trying to help. ISTM that if
> Robert's concern is not addressed one way or another, you will just
> get "rejected" on this basis.
>
Yes, I understand you are trying to help, and I appreciate it! My
opinion, and that of others as well from the original thread, is that
this patch should either go in as is and break that one case, or not go
in at all. I'm fine with either (although clearly I would prefer it
went in otherwise I wouldn't have written the patch).
--
Vik