[tor-dev] Anyone wanting to write some Weather-tight code?

[Karsten Loesing]

On 1/10/14 3:42 PM, Abhiram Chintangal wrote:
> On 01/09/14 22:07, Damian Johnson wrote:
>>> Hello Karsten,
>>>
>>> It looks like a nice little project to work on. I reguarly check my
>>> relay-status via Atlas, at the moment using the Onionoo service makes sense
>>> to me too.
>>>
>>> For now the onionoo-glue-code, seems like a good place to start.  Do you
>>> have any other suggestions or ideas?
>>>
>>> thanks!
> 
>> Hi Abhiram, glad you want to tackle this! I'm not sure if it makes
>> sense with the plan to rewrite Weather but six months back I submitted
>> a patch to swap the present Weather from TorCtl to Stem...
>>
>> https://trac.torproject.org/8264
> 
> Sounds great, I took a cursory look at the Django application in the
> repo, I should be able to reuse many modules from there.

Great!

>> Testing and merging that might be a fine way of getting acquainted to
>> the present codebase and functionality.
> 
> If I understood everything so far, I should look at ways to use
> Onionoo's restful service instead of the current approach which
> uses stem to grab hourly consensus updates of via the tor-network.

At least that's my suggestion, and I think it'll save us a lot of work
in the long run.  When I designed Onionoo, I had Weather in mind as
possible client application.  I think that Weather shouldn't have to
implement its own Tor network status database.  That's something that
Onionoo already does, and it was painful enough to write.  No reason to
maintain quite similar code in Weather.

Obviously, there are also reasons against using Onionoo for the Weather
rewrite.  For example, fixing current Weather *might* take less effort
than replacing its status database with an Onionoo client.  And the new
Weather will depend on the Onionoo service, rather than on a larger set
of Tor directory authorities.

Please don't hesitate to add your own pros and cons to this list.

All the best,
Karsten