[Evolution] Bug#587302: Bug#587302: Bug#587302: evolution: Forced offline by error

[Christophe Lohr]

Le 27/06/2010 17:07, Yves-Alexis Perez a écrit :
>> But Evo seems to check if the "global system is online" (why?)
>>      
> Because you installed network-manager.
>    

I'm convince that it is out of it's scope.
Evo just need to reach email servers, and nothing else.
In some circumstances, these servers may be local, on an offline system...
At the opposite, servers may be unreachable even if the system is said 
to be online (according to NM semantics: one of the interfaces it has to 
manage is up)

The only way to know if servers are reachable is to try a connections.
Do one need to check it before all?

> It's not incompatible. You just seem to not use network-manager the way
> it's intended to.
>    

Maybe.

However, are you sure that Evo use it the way it's intended to?

>> I use NM to manage the wifi interface. It does well the job. For the 3G
>> interface, I have some scripts, out of the scope of NM. From time to
>> time, I also have a VPN, not managed by NM.
>> So, it's not a good idea to ask NM if I am online or not.
>>      
> If you don't like NM to be able to give a decision on online/offline
> status, don't use NM.

I like NM, it does a good job for managing interfaces it is configured to.
Moreover, its decision on online/offline status is relevant, except that 
it concerns only interfaces it is configured to manage. Developers of 
other applications may misunderstand this point and use NM without 
precautions.

> *Now*, I can understand that some people would like to, from time to
> time, override offline decision. That may be a bug, but your best bet is
> to open a bug *upstream* instead of here.
>    

According to your suggestion, I posted this:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=587325

Many thanks for your feed back.
Regards