[Sugar-devel] [DESIGN] Rework the wireless control panel section to respect non-wireless connections

[Frederick Grose]

On Wed, Jan 22, 2014 at 2:56 PM, Frederick Grose <fgrose at gmail.com> wrote:

> On Wed, Jan 22, 2014 at 6:03 AM, Daniel Narvaez <dwnarvaez at gmail.com>wrote:
>
>> On 21 January 2014 03:49, Frederick Grose <fgrose at gmail.com> wrote:
>>
>>> On Mon, Jan 20, 2014 at 4:17 PM, Frederick Grose <fgrose at gmail.com>wrote:
>>>
>>>> Fixes #4410  https://bugs.sugarlabs.org/ticket/4410
>>>>
>>>> The 'Discard network history' button is peculiar to Sugar, and
>>>> is destructive to all non-wireless connections, despite being in
>>>> the 'Wireless' section of the control panel.
>>>>
>>>> In the proposed patch[0], the code variables and functions are
>>>> renamed and coded to be specific to wireless connections.
>>>>
>>>> The 'Radio' checkbox instruction and 'Discard ...' button are
>>>> reworded to avoid confusion over their state and actions.[1]
>>>>
>>>> If wireless hardware is not present on the system, the 'Wireless'
>>>> section of the control panel is excluded.[2]
>>>>
>>>> [0,1,2] - https://github.com/sugarlabs/sugar/pull/228
>>>>
>>>
>>> Daniel Narvaez
>>> wrote via
>>>  notifications at github.com
>>>
>>>> It seems like the code is handling adhoc networks specially, but I
>>>> didn't find an explanation of that change. Can you please elaborate on it?
>>>
>>>
>>> The current code at,
>>>
>>> https://github.com/sugarlabs/sugar/blob/master/src/jarabe/model/network.py#L835
>>>
>>> tests for sugar-internal connections by reading the connection_id string.
>>>
>>> The new code is specific to wireless connections.
>>> The first condition, the presence of a wifi_settings list, avoids 7 of
>>> the 7 sugar-internal
>>> connections on XO-1.  The sugar ad hoc connections are avoided by a
>>> second condition,
>>> the same test as in the currrent code,
>>> connection_id.startswith(ADHOC_CONNECTION_ID_PREFIX).
>>>
>>
>> So to make that more explicit, compared to the old code we are *not*
>> clearing anymore gsm, mesh and xs_mesh networks. Is that correct?
>>
>
> The sugar-internal gsm, mesh, & xs_mesh connections are not cleared in the
> old or proposed, new code.
>
>> Do we have another way to clear gsm?
>>
>
> There is a separate control panel for GSM modems. I'm not very familiar
> with it though.
>
> I'm not too sure we should stop clearing mesh and xs_mesh. It is wifi too
>> and it might play a role in the unbreak procedures people have describing
>> in this thread.
>>
>
> See above (the sugar-internal mesh and gsm are not cleared).   I suppose
> someone might create another such connection at a command line.
>
>>  What's the harm of doing that?
>>
>
> Clearing non-sugar-internal mesh connections should be ok (as that is the
> current behavior).
>
> GSM matters might better be handled on its own control panel.
>
> So yes, GSM users creating custom, non-sugar-gsm connections would be
> affected if they depended on the 'Discard network history' button.
> The modemconfiguration panel, though, seems to use only the one
> GSM_CONNECTION_ID = 'Sugar Modem Connection'
> from jarabe.model.network.
> So perhaps best for now, to continue to discard non-sugar GSM connections,
> and test for other GSM connections before excluding that control panel
> subsection.
>

This seems too much of a corner case (along with other manually created,
non-wireless connections).

>
>
> I'll close the pull request and resubmit changes sequentially as
> recommended in the other note.
>
> Thanks for reviewing,          --Fred
>

There is a new pull request at
https://github.com/sugarlabs/sugar/pull/234

And a feature page here:
http://wiki.sugarlabs.org/go/Features/Network_control_panel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sugarlabs.org/archive/sugar-devel/attachments/20140128/0ee68170/attachment-0001.html>