Re: [bug fix] Savepoint-related statements terminates connection

On 1 September 2017 at 08:09, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote:
> On Fri, Sep 1, 2017 at 3:05 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
>> I'm not sure I see the use case for anyone using SAVEPOINTs in this
>> context, so simply throwing a good error message is enough.
>>
>> Clearly nobody is using this, so lets just lock the door. I don't
>> think fiddling with the transaction block state machine is anything
>> anybody wants to do in back branches, at least without a better reason
>> than this.
>
> I don't think you can say that, per se the following recent report:
> https://www.postgresql.org/message-id/CAH2-V61vxNEnTfj2V-zd+mA-g6kQMJgd5SvXoU3JBvdzQH0Yfw@mail.gmail.com

AIUI, nobody is saying this should work, we're just discussing how to
produce an error message. We should fix it, but not spend loads of
time on it.

I've added tests to the recent patch to show it works.

Any objection to me backpatching this, please say.

--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services