Getting the API server URL in workers

[Andrew Wilkins]

On Wed, Sep 3, 2014 at 10:49 PM, John Meinel <john at arbash-meinel.com> wrote:

> It feels like this should be something that the API server is looking up
> in the database/some sort of query, not just munging a URL. I'm fine with
> having a Params struct, though I worry that it isn't really appropriate to
> be passing all possible Params to all facades. They already have a State
> connection, I'd like us to at least be designing the layering
> appropriately.  I feel like state.State is intended to be where "Knowledge"
> resides. Why isn't what we want there?
>
It is in there, it just seems a bit off to go to State when we have the
information already. OTOH, it is nice that the facades are mostly ignorant
of their container. I'll investigate constructing the URL from State.

>  John
> =:->
> On Sep 3, 2014 6:03 PM, "Andrew Wilkins" <andrew.wilkins at canonical.com>
> wrote:
>
>> I'm working on moving all tools downloads to download from the API
>> server. There will are a few APIs that return tools:
>>   - Upgrader.Tools
>>   - Client.FindTools
>>   - Provisioner.FindTools
>>
>> These APIs will need to return URLs pointing at the API server. I'm
>> intending to change the facade constructors to take a parameter struct, and
>> extend it with the API server's root URL (i.e. https://<address>:<port>/environment/<uuid>),
>> where the address is the one used by the client to connect.
>>
>> Any reason I should not go ahead and do that? This will probably make it
>> easier to slot in a "Restarter" or whatever as well.
>>
>> Cheers,
>> Andrew
>>
>> --
>> Juju-dev mailing list
>> Juju-dev at lists.ubuntu.com
>> Modify settings or unsubscribe at:
>> https://lists.ubuntu.com/mailman/listinfo/juju-dev
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju-dev/attachments/20140904/34ad6b9e/attachment.html>