I wrote:
> Hmm, excellent point. While our code response to all these errors
> should be the same, you are right that that doesn't extend to emitting
> identical error texts. For EHOSTUNREACH/ENETDOWN/ENETUNREACH, we
> should say something like "connection to server lost", without claiming
> that
Fujii Masao writes:
> On 2020/10/09 4:15, Tom Lane wrote:
>> -#ifdef ECONNRESET
>> -case ECONNRESET:
>> +case ALL_CONNECTION_LOSS_ERRNOS:
>> printfPQExpBuffer(&conn->errorMessage,
>>
On 2020/10/09 4:15, Tom Lane wrote:
Over in the thread at [1], we've tentatively determined that the
reason buildfarm member lorikeet is currently failing is that its
network stack returns ECONNABORTED for (some?) connection failures,
whereas our code is only expecting ECONNRESET. Fujii Masao
At Thu, 08 Oct 2020 21:41:55 -0400, Tom Lane wrote in
> Kyotaro Horiguchi writes:
> > At Thu, 08 Oct 2020 15:15:54 -0400, Tom Lane wrote in
> >> Accordingly, I propose the attached patch (an expansion of
> >> Fujii-san's) that causes us to test for all five errnos anyplace
> >> we had been che
Kyotaro Horiguchi writes:
> At Thu, 08 Oct 2020 15:15:54 -0400, Tom Lane wrote in
>> Accordingly, I propose the attached patch (an expansion of
>> Fujii-san's) that causes us to test for all five errnos anyplace
>> we had been checking for ECONNRESET.
> +1 for the direction.
> In terms of conn
At Thu, 08 Oct 2020 15:15:54 -0400, Tom Lane wrote in
> Over in the thread at [1], we've tentatively determined that the
> reason buildfarm member lorikeet is currently failing is that its
> network stack returns ECONNABORTED for (some?) connection failures,
> whereas our code is only expecting E
Over in the thread at [1], we've tentatively determined that the
reason buildfarm member lorikeet is currently failing is that its
network stack returns ECONNABORTED for (some?) connection failures,
whereas our code is only expecting ECONNRESET. Fujii Masao therefore
proposes that we treat ECONNAB
