Re: Suggestion: Issue warning when calling SET TRANSACTION outside transaction block

Robert Haas wrote
> On Mon, Nov 18, 2013 at 9:07 PM, Bruce Momjian <

> bruce@

> > wrote:
>> Well, ERROR is what LOCK returns, so if we change SET TRANSACTION to be
>> WARNING, we should change LOCK too, so on backward-compatibility
>> grounds, ERROR makes more sense.
>>
>> Personally, I am fine with changing them all to WARNING.
> 
> I don't think it's worth breaking backward compatibility.  I'm not
> entirely sure what I would have decided here in a vacuum, but at this
> point existing precedent seems determinative.

Well, at this point we have already broken backward compatibility by
releasing this.  With Tom's thread necromancy I missed the fact this got
released in 9.3

Now, given normal upgrade realities the people likely to have this bite them
probably are a ways out from upgrading so I wouldn't expect to have seen
many complaints yet - but at the same time I do not recall seeing any
complaints yet (limited to -bugs and -general)

The referenced patch:

is released
is documented
is consistent with precedent established by similar codepaths
causes an obvious error in what is considered broken code
can be trivially corrected by a user willing and able to update their
application

I'd say leave this as-is and only re-evaluate the decision if complaints are
brought forth.

David J.




--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Suggestion-Issue-warning-when-calling-SET-TRANSACTION-outside-transaction-block-tp5743139p5779170.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.