Qingqing Zhou <zhouqq(dot)postgres(at)gmail(dot)com> writes:
> I do feel that we have too many functions instructing how to handle
> interrupts and they are subtle - I just found a new friend
> HOLD_CANCEL_INTERRUPTS :-(
Indeed, which is why I think a patch for this issue should not introduce
a new mode/context in which proc_exit() is executed. The code path would
be so seldom executed that we'd never manage to isolate any bugs that
might be in it. Thus also my objection to Alvaro's patch that had us
going to proc_exit from partway through the error cleanup sequence.
regards, tom lane