On Tue, Jul 16, 2013 at 2:14 PM, Kris Borchers <kris(a)redhat.com> wrote:
On Jul 16, 2013, at 7:10 AM, Matthias Wessendorf <matzew(a)apache.org>
wrote:
On Tue, Jul 16, 2013 at 2:05 PM, Kris Borchers <kris(a)redhat.com> wrote:
>
> On Jul 16, 2013, at 3:05 AM, Matthias Wessendorf <matzew(a)apache.org>
> wrote:
>
>
>
>
> On Tue, Jul 16, 2013 at 9:57 AM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
>
>>
>>
>>
>> On Tue, Jul 16, 2013 at 8:36 AM, Matthias Wessendorf
<matzew(a)apache.org>wrote:
>>
>>>
>>>
>>>
>>> On Mon, Jul 15, 2013 at 8:34 PM, Lucas Holmquist
<lholmqui(a)redhat.com>wrote:
>>>
>>>> here's a scenario.
>>>>
>>>> A user creates an "Application" then adds a new "iOS
Variant" to it.
>>>>
>>>> they decide they need to rename this Variant since they didn't name
it
>>>> very well.
>>>>
>>>> There is currently no way of renaming/ updating the description of
>>>> this variant without re-sending the Certificate/pass phrase. this would
be
>>>> the same for Android and Simple Push, just substitute google key and
>>>> channels
>>>>
>>>
>>> yes - but should be easy to change :-)
>>>
>>> See here:
>>>
>>>
https://github.com/aerogear/aerogear-unified-push-server/blob/master/src/...
>>>
>>> in the "update", we just need to onbly update/replace those things
that
>>> are NOT null. Of course, perhaps we need to rethink the "update
validation"
>>> too:
>>>
>>>
https://github.com/aerogear/aerogear-unified-push-server/blob/master/src/...
>>>
>>> Yes that could work even if I must admit I'm not a big fan of updating
>> based on a null check ... Imagine someone wants to delete the description
>> by setting it to null , that won't be possible ;)
>>
>
> well, yeah - that (the desc) is optional - I was more interested in the
> required fields:
> e.g. for iOS Variant:
> * cert/passphrase (or googlekey/networkURL on other variants)
> PushApp:
> * name
>
>
> I'm not sure how I feel about this partial update. I'm not a big fan of
> saying "It's a required field but leave it null if you don't want it to
> change." Feels wrong.
>
not sure I am following, but the required fields should never be null
Now I'm not following.
:-)
Let's go through a scenario and see how each would work with your
method.
Imagine there are currently values in every field. What fields need values
and what can be null when:
I update a single required field?
I update a single optional field?
Well,when I update the FORM and hit ENTER (let's assume this is a PUT, just
for now..), aren't the entire form values sent?
So, any field that changed will be udpated, I think
Now if you NULL out a required field, I'd imagine that is not allowed.
>
> If we are to do partial updates like this though, I would say it should
> probably use http PATCH instead of PUT, right?
>
>
>
>
>
>
>> (Well not sure it is a valid use case but you see my point) and as you
>> said we must then rethink user validation.
>> Other option is to split the update into more smaller specific updates
>> methods ... but not sure about this also.
>>
>
>
>
>
>
>>
>>
>>
>>>
>>>
>>>>
>>>> Also, once we have a variant created, we have a variant ID, we
>>>> should be able to use that ID( plus type? ) to do updates instead of
having
>>>> to pass the applicaitonID and variantID( same would go for a delete i
guess
>>>> ).
>>>>
>>>>
>>>> Not really sure the best way to handle this.
>>>>
>>>
>>> the "variantID" is not a PK, but it's unique - so a new finder
is what
>>> you need.
>>>
>>>
>>>
>>>>
>>>> Should a user be able to update a certificate/key/channel once it has
>>>> been created.
>>>
>>>
>>> yes, they should be able to replace an outdated (or revoked)
>>> certificate, API etc.
>>>
>>>
>>>
>>>> How would this affect any installations it might have?
>>>>
>>>
>>> Not at all. so if (speaking Androind) ... a new "project" is used
-
>>> there is a new SenderID - that's more impact, since the actual mobile
app
>>> needs to be updated, but again that's not really a concern for the push
>>> server
>>>
>>>
>>>
>>>>
>>>> thoughts?
>>>>
>>>> -Luke
>>>> _______________________________________________
>>>> aerogear-dev mailing list
>>>> aerogear-dev(a)lists.jboss.org
>>>>
https://lists.jboss.org/mailman/listinfo/aerogear-dev
>>>>
>>>
>>>
>>>
>>> --
>>> Matthias Wessendorf
>>>
>>> blog:
http://matthiaswessendorf.wordpress.com/
>>> sessions:
http://www.slideshare.net/mwessendorf
>>> twitter:
http://twitter.com/mwessendorf
>>>
>>> _______________________________________________
>>> aerogear-dev mailing list
>>> aerogear-dev(a)lists.jboss.org
>>>
https://lists.jboss.org/mailman/listinfo/aerogear-dev
>>>
>>
>>
>> _______________________________________________
>> aerogear-dev mailing list
>> aerogear-dev(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/aerogear-dev
>>
>
>
>
> --
> Matthias Wessendorf
>
> blog:
http://matthiaswessendorf.wordpress.com/
> sessions:
http://www.slideshare.net/mwessendorf
> twitter:
http://twitter.com/mwessendorf
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
>
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
--
Matthias Wessendorf
blog:
http://matthiaswessendorf.wordpress.com/
sessions:
http://www.slideshare.net/mwessendorf
twitter:
http://twitter.com/mwessendorf
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
--
Matthias Wessendorf
blog:
http://matthiaswessendorf.wordpress.com/
sessions:
http://www.slideshare.net/mwessendorf
twitter:
http://twitter.com/mwessendorf