[ntp:hackers] Cool new stuff

David L. Mills mills at udel.edu
Mon Jul 24 03:32:46 UTC 2006


Danny,

I hear you and am equally annoyed about C reserved words. However, we 
would really like to separate the token parsing and the syntactic 
analysis for the same reasons that C does. We also require the syntax to 
be an unambiguousd, one-lookahead, context-free grammar, again for the 
same reasons that C does.

Dave

Danny Mayer wrote:

> Sachin Kamboj wrote:
>
>> Dear Prof. Mills, Fellow hackers,
>>
>> I will need some volunteers and testers to help with the testing. I have
>> tried to exercise all parts of the configuration code but there might be
>> subtle bugs on different platforms that I don't have access to. I 
>> will fix
>> these as they are reported.
>>
>> Also, I have added a couple of source files that are being included into
>> ntp_config.c (yuck!). Ideally they should be compiled seperately and 
>> linked
>> with the rest of the code. I will need help adding that to the configure
>> procedure.
>>
>> There are two main departures in the configuration syntax from the 
>> existing
>> practise:
>> 1. peerstats, loopstats, cryptostats, sysstats, clockstats and 
>> rawstats are
>> now reserved words and may not be used as filenames in the configuration
>> file. Hence, a configuration command such as:
>>
>> filegen peerstats file peerstats enable
>>
>> won't work and will give a syntax error. Instead, use something like:
>>
>> filegen peerstats file peer_stats enable
>>
>
> I'm not sure I understand why they need to be reserved words. The syntax
> should be clearly established to differentiate between a name and a
> value on a line. If you can't then the syntax should be revised.
>
>> 2. Right now, the -4 and -6 switches, for DNS zone selection don't 
>> work. I
>> wasn't sure why they were needed, but I can put them in later.
>>
> They are needed because we run over both IPv4 and IPv6 and DNS lookups
> will give you back both addresses. Sometimes you don't want the IPv6
> addresses because your network infrastructure. Sometimes, and this is
> more for the future, you only want to deal with IPv6 infrastructures.
>
>> Other comments:
>>
>
>> Also if anyone has some thoughts on why everything works fine when 
>> ntpd is
>> run with the -d switch, but doesn't configure when run without it, please
>> let me know. That is the final block preventing me from releasing the 
>> code
>> into the wild.
>>
>
> That one's obvious. You have something defined within a if(debug) that
> shouldn't be conditional. Since the code is not yet released, we can
> tell you where.
>
> Danny
>
>> Thanks for any suggestions/comments.
>>
>> Regards,
>> Sachin.
>>
>



More information about the hackers mailing list