Hi,
On 2020-03-23 13:54:31 -0400, Tom Lane wrote:
> Andres Freund writes:
> > On 2020-03-22 18:30:04 -0400, Tom Lane wrote:
> >> Yeah, I was wondering about giving that a new result code, too.
> >> It would be a little bit invasive and not at all back-patchable,
> >> but (say)
Andres Freund writes:
> On 2020-03-22 18:30:04 -0400, Tom Lane wrote:
>> Yeah, I was wondering about giving that a new result code, too.
>> It would be a little bit invasive and not at all back-patchable,
>> but (say) TM_SerializationViolation seems like a cleaner output ---
>> and we could
Hi,
On 2020-03-22 18:30:04 -0400, Tom Lane wrote:
> Andres Freund writes:
> > I wonder if we shouldn't just change the crosscheck case to set
> > something other than TM_Updated, as it's not really accurate to say the
> > tuple was updated.
>
> Yeah, I was wondering about giving that a new
Andres Freund writes:
> On 2020-03-22 15:00:51 -0400, Tom Lane wrote:
>> Another thing that is very peculiar in this area is that the initial
>> assertion in the second stanza allows the case of result == TM_Deleted.
> In this case, isn't it clearly required to accept TM_Deleted? The HTSU
>
Hi,
On 2020-03-22 15:00:51 -0400, Tom Lane wrote:
> Antonin Houska writes:
> > I was trying to figure out what exactly the "crosscheck snapshot" does in
> > the
> > RI checks, and hit some assertion failures:
>
> Yeah, your example reproduces for me.
>
> > I'm not familiar enough with this
Antonin Houska writes:
> I was trying to figure out what exactly the "crosscheck snapshot" does in the
> RI checks, and hit some assertion failures:
Yeah, your example reproduces for me.
> I'm not familiar enough with this code but I wonder if it's only about
> incorrect assertions.
Mmm ...
