[Nut-upsdev] [nut-commits] svn commit r1846 - in trunk: . clients common drivers include man server

[Arjen de Korte]

Citeren Arnaud Quette <aquette.dev op gmail.com>:

>> Hmm.. won't relative timestamps make it difficult when you're debugging
>> program interaction? eg a UPS driver & upsd.
> yup, that was the core of my last reply... (an option to override the
> elapsed time and switch to a timestamp system).

I don't think this is needed.

> Arjen's version (elapsed time) is very nice for working on 1 component,
> while the timestamp would help more in cross components (program
> interaction). possibly using the new "-T" to switch to the timestamp mode...

The socket/network protocol between the separate programs is simple  
and entirely in clear text. So it isn't hard to follow what is going  
on when looking at just one end of the connection, provided the debug  
level is high enough and both what is sent and received is being logged.

It has been like this for years and we have never really needed to run  
both ends in debug mode when trying to diagnose what is going on  
(remember, so far we didn't have any time stamps at all). I understand  
what the objections may be, but I really doubt this will add anything  
useful (the recent logs didn't hint in that direction either).

Best regards, Arjen
-- 
Please keep list traffic on the list