[Nodeshot] Latest changes thanks to your feedback

[Clauz]

On 11/24/2013 04:21 PM, Nemesis wrote:
> On 11/20/2013 09:25 PM, Clauz wrote:
>> On 11/20/2013 06:22 PM, Nemesis wrote:
>>> Hi,
>>>
>>> First of all, thankyou all for the feedback you are sending.
>>>
>>> I've done most of thechanges that I needed to do after receiving your
>>> answers to my latest technical questions.
>>> This is surely the way to go, after receiving yourfeedback I feel much
>>> more secure, so please keep to do so!
>>>
>>> And good news, hey Clauz, look at this code:
>>>
>>> *>>> device = Device.objects.get(name='RM5PomeziaSNode')**
>>> *(0.015) SELECT "net_device"."id", "net_device"."added",
>>> "net_device"."updated", "net_device"."access_level",
>>> "net_device"."name", "net_device"."node_id", "net_
>>> device"."type", "net_device"."status", "net_device"."location",
>>> "net_device"."elev", "net_device"."os", "net_device"."os_version",
>>> "net_device"."first_seen", "
>>> net_device"."last_seen", "net_device"."description",
>>> "net_device"."notes", "net_device"."data", "net_device"."shortcuts" FROM
>>> "net_device" WHERE "net_device"."
>>> name" = 'RM5PomeziaSNode' ; args=('RM5PomeziaSNode',)
>>> *>>> device.connector.connect()**
>>> *True
>>> *>>> device.connector.get_os()**
>>> *('AirOS', 'XM.ar7240.v5.3.3.sdk.9634.111221.2238')
>>> *>>> device.connector.get_wireless_output_power()*
>>> 27
>>> *>>> device.connector.get_wireless_channel()**
>>> *'5765'
>>>>>> *device.connector.disconnect()*
>>> I'm slowly getting the syntactic sugar I desire! And it's self
>>> explanatory! :)
>> Very sweet! :)
>>
>>> It willbe super cool to have a connector_class that uses SNMP or any
>>> other lighter protocol instead of SSH.
>>> This would save us the need of having to call "connect()" and
>>> "disconnect()".
>> Perhaps you can connect on the __init__ of the connector object and
>> disconnect on __del__ ?
>>
> 
> Hey clauz,

Hi.

> I found a solution:
> if the instance has to execute a command but is not yet connected it
> will connects automatically before proceeding.
> The connection should be closed automatically by the __del__ method as
> you suggested, although the python docs say this method is not
> guaranteed to be called, for that reason in the internals I'm still
> using connect() and disconnect(), just to be sure.
> 
> Updated example:
> *>>> pomezia = Device.objects.get(name='RM5PomeziaSNode')*
> (0.016) SELECT "net_device"."id", "net_device"."added",
> "net_device"."updated", "net_device"."access_level",
> "net_device"."name", "net_device"."node_id", "net_
> device"."type", "net_device"."status", "net_device"."location",
> "net_device"."elev", "net_device"."os", "net_device"."os_version",
> "net_device"."first_seen", "
> net_device"."last_seen", "net_device"."description",
> "net_device"."notes", "net_device"."data", "net_device"."shortcuts" FROM
> "net_device" WHERE "net_device"."
> name" = 'RM5PomeziaSNode' ; args=('RM5PomeziaSNode',)
> *>>> pomezia.connector.get_os()**
> **('AirOS', 'XM.ar7240.v5.3.3.sdk.9634.111221.2238')*
>>>>
> *>>> pomezia.connector.get_wireless_dbm()**
> **'-74'**
> *>>>
> *>>> pomezia.connector.get_wireless_mode()**
> **'ap'**
> *

I didn't know that the __del__ method was not guaranteed to be called. I
like this approach.

> I also moved the connector classes into the "ssh" folder, so it's
> explicit that for other protocols the other classes will need to be in
> an "snmp" folder.
> It would be cool to experiment with an snmp connector prototype.. if we
> lay it down together I should be able to go ahead alone.
> Shall we organize an evening somewhen in the next month (before
> christmas)? Like a Monday, Tuesday or Friday for me it's ok

Yes, let's try to schedule it. Thursday during the ninux Rome weekly
meeting is bad?

Clauz