[Dnsmasq-discuss] [PATCH 3] Some upstream replies not being logged

[Simon Kelley]

On 01/09/2021 12:45, Dominik DL6ER wrote:
> Hey Simon,
> 
> On Tue, 2021-08-31 at 21:09 +0100, Simon Kelley wrote:
>> I've tagged 2.86rc2. I've systematically tested the affected code,
>> and we're dog-fooding it now. It would be good to get as much other
>> testing in as possible before the 2.86.
> 
> We have automated CI testing putting a lot of queries (and different
> query types) through it and nothing unexpected happened.
> We've added the patch into our beta testing so we have at least a few
> dozen additional testers of the new code on our side, too.
> 
> I've seen two minor things but this is really nit-picky:
> 
>> dig TYPE65 https.dns.netmeister.org
> 
> leads to
> 
>> reply https.dns.netmeister.org is type=65
> 
> whereas I'd consider
> 
>> reply https.dns.netmeister.org is [type=65]
> 
> slightly more elegant because it is the type of the reply, not the
> content. The necessary change would be querystr(NULL, aqtype) ->
> querystr("", aqtype) in the "else" branch of "if (aqtype == T_TXT)" but
> that's entirely your call.

For consistency if we know the type we do

reply https.dns.netmeister.org is <TXT>

so if we don't know the type we should do

reply https.dns.netmeister.org is <type=65>

I've implemented that.

> 
> Also, DNSSEC signed TXT records were not logged as such with log-
> queries=extra. print_txt() needs secflag for this. Patch attached.
> 

Patch applied. Removing the declaration for int i breaks when building
DNSSEC code, so I changed that to use the j counter variable instead.

> I'll report back if anything odd comes up. Silence will be a good
> thing.

Great, thanks for your help. It's all running fine here too.

I tagged 2.86rc3 for completeness.

Simon.


> 
> Best,
> Dominik
>