[ntp:hackers] new config parsing code

David L. Mills mills at udel.edu
Wed Apr 25 10:31:18 PDT 2007


Brian,

The simulator is only an interface to the actual daemon and is tightly 
integrated in the source code. It is enabled at compile time by a 
confuguration keyword. This is the same case as with the current NTP 
distribution.

The simulator tool has been a vital component in the past for debugging 
things impractical to test in vivo. The actual code between the ifdefs 
could be removed, but this would raise serious problems in maintenance 
and support and result in two streams of bugfixes. Better to leave it in 
and never divulge the enabling configuration keyword.

Dave

Brian Utterback wrote:

> Is the simulator something that I would want to ship to my customers,
> or something I use to verify and/or diagnose NTP's behavior on a
> particular OS rev? If so, I might be inclined to pass it around 
> internally at Sun, but not to ship in to customers. If I am wrong,
> then help me to understand the customer usage of the simulator.
> 
> 
> David L. Mills wrote:
> 
>> Brian,
>>
>> So far as I am concenred, the simulator is an integral component in 
>> the distribution, especially as it has enhanced functionality over the 
>> previous version. As you say Sun will not include it in their 
>> distribution, I assume Sun will commit such resources as to maintain a 
>> separate corporate distibution. In that case, I assume Sun will commit 
>> such resources to aid in the related development effort.
>>
>> You should be reminded that Digital did commit a warm body to aid in 
>> the development of NTPv3.
>>
>> Dave
>>
>> Brian Utterback wrote:
>>
>>> Harlan Stenn wrote:
>>>
>>>> Folks,
>>>>
>>>> Sachin has almost finished the beast.  To our knowledge the last issue
>>>> that needs to be resolved is a linker problem building ntpdsim.  (There
>>>> are functions defined in libntp.a that we want to override for ntpdsim
>>>> by creating copies in a .o file, and the linker is seeing both and
>>>> complaining instead of using the versions in the .o file and ignoring
>>>> the later definitions in the .a file.)
>>>>
>>>> I can either commit the code now and we work on fixing this linker 
>>>> issue
>>>> ASAP, or I can wait until we resolve the link problem before committing
>>>> the code.
>>>>
>>>> Any preferences out there?
>>>>
>>>> H
>>>> _______________________________________________
>>>> hackers mailing list
>>>> hackers at support.ntp.org
>>>> https://support.ntp.org/mailman/listinfo/hackers
>>>>   
>>>
>>> I vote for sooner, but I have a vested interest in getting it done 
>>> soon and have no such
>>> interest in the simulator since we will not be delivering it.
>>
>>
>>
>> _______________________________________________
>> hackers mailing list
>> hackers at support.ntp.org
>> https://support.ntp.org/mailman/listinfo/hackers
> 
> 


More information about the hackers mailing list