Call for inncheck testing

Julien ÉLIE julien at trigofacile.com
Fri Jul 8 20:25:42 UTC 2011


Hi Florian,

> picking up from news.software.nntp, what do people here think about
> uppercase vs lowercase letters in IPv6 addresses? I tend to think we
> should be permissive

So do I.
Uppercase characters must be accepted, as you say:

>> RFC 5952 section 4 / 4.3 says that all characters in an IPv6 address
>> MUST be textually represented in lowercase and advises that humans
>> should do so, too.  At the same time, all implementations must be able to
>> accept both lowercase and uppercase.



>>> /var/lib/news/etc/readers.conf:10: not a valid option name: require_ssl
>>> /var/lib/news/etc/readers.conf:10: option true must be immediately followed
>>> by a colon
>>> /var/lib/news/etc/readers.conf:10: not a valid option name: true
>>
>> BTW, do people think that the second and third line are ugly and
>> misleading? I just thought of a small fix that would suppress further
>> error messages until the parser is back in sync, at the danger of hiding
>> two consecutive errors until the first one gets fixed...
>
> I *do* find it ugly and misleading, and here's what I thought to do
> about it

That is much better for the user.  Thanks for the improvement.



> I'd think it's clearer to write
>
>      size: 16384,
>
> to indicate that it means "16384 and above", rather than without the
> comma. And it's the same for mintime/maxtime, although that's hardly
> ever used, no?

Hardly is not never :-)



>> The "include" directive is not understood. (Be careful that both
>> readers.conf and included files can be partially formatted files.
>> It means that readers2.conf can for instance consist only of a
>> closing bracket "}"???)
>
> No. Admins are free to break and obfuscate their configuration in
> all the ways they want, but that's for the "new" parser to sort out
> then. To keep things manageable, I'm not going to hunt for included
> files, and require that what gets included is a "complete" set of
> blocks and options. That is to say, inncheck is going to ignore the
> include statement, and if the admin wants to check the included file
> he has to run inncheck again with readers.conf=includefile. Same for
> innfeed.conf's $INCLUDE. Does that sound sensible?

Totally sensible.
It works for me.


All of your patches about the subjects dealt with this mail are 
committed.  Thangs again, Florian.

-- 
Julien ÉLIE

« Oublie les injures, n'oublie jamais les bienfaits. »



More information about the inn-workers mailing list