Re: pgsql: Fix (some of the) breakage introduced into query-cancel

Simon Riggs <simon@2ndQuadrant.com> writes:
> Ah, you were referring to those lines. I thought you were referring to
> the additional lines you put in around ClientAuthInProgress.

That's not really additional code --- I was just moving (and duplicating
:-() some of the steps so that they wouldn't get executed if we choose
to fall out without processing the query cancel request.  The previous
logic here was essentially to prevent ProcessInterrupts from being
called at all while DoingCommandRead.  The new logic is to call it and
have it understand that it shouldn't do anything if DoingCommandRead
(except when the HS state indicates an override).

            regards, tom lane