james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From spfh...@caseyconnor.org
Subject Re: jSPF nit
Date Wed, 26 Aug 2009 21:03:27 GMT
Opened as JSPF-83... I only had/have time to submit a kind of paltry 
recommendation, but hopefully it helps. Thanks!
-Casey

Robert Burrell Donkin wrote:
> On Thu, Aug 20, 2009 at 11:49 PM, <spfhelp@caseyconnor.org> wrote:
>   
>> Hi - maybe this is on the picky side, feel free to say so, but:
>>
>> org.apache.james.jspf.impl.DNSServiceXBillImpl.java - line 158-160, at least
>> in rev 801912:
>>
>>           if (queryResult == Lookup.TRY_AGAIN) {
>>               throw new TimeoutException(query.getErrorString());
>>           }
>>
>> ...i'm basically ignorant of the finer points of all this, but doesn't
>> org.xbill.DNS.Lookup return TRY_AGAIN for a variety of reasons (e.g. not
>> just Timeouts) which would deserve potentially different exceptions/logging?
>> from Lookup.java:
>>
>>  531            if (badresponse) {
>>  532                    result = TRY_AGAIN;
>>  533                    error = badresponse_error;
>>  534                    done = true;
>>  535            } else if (timedout) {
>>  536                    result = TRY_AGAIN;
>>  537                    error = "timed out";
>>  538                    done = true;
>>  539            } else if (networkerror) {
>>  540                    result = TRY_AGAIN;
>>  541                    error = "network error";
>>  542                    done = true;
>>
>>
>> ...we were seeing "timeouts" from jSPF in our logs that were happening in
>> <20ms, so we believe we're having other net issues in reality.
>>     
>
> sounds about right - could you open a JIRA (issues.apache.org/jira)
> and contribute a suggested patch?
>
> cheers
>
> - robert
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
> For additional commands, e-mail: server-dev-help@james.apache.org
>
>   

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message