10:17 a.m.
Hi,
I went once again through
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
says that Sender API should be fire&forget. It feels more like "maybe
fire"&forget, for instance it does not say that your credentials were wrong - or
it says, you need parse logs to get that information.
If I think about Android, iOS, JS solutions to communicate with
UnifiedPush we provide - Pipes - they always provide a callback to be executed
on success/failure. Could we add callback to Sender API? Or should not Aerogear
rather have something like Pipes abstraction for Java developers instead of
pretty dumb Sender API?
Thoughts?
Thanks,
Karel
10:22 a.m.
There server does send a response, "submitted" or something like that. So maybe
there should be callbacks
the node.js sender has callbacks for success and error
On Sep 17, 2013, at 11:17 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
Hi,
I went once again through
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
says that Sender API should be fire&forget. It feels more like "maybe
fire"&forget, for instance it does not say that your credentials were wrong -
or
it says, you need parse logs to get that information.
If I think about Android, iOS, JS solutions to communicate with
UnifiedPush we provide - Pipes - they always provide a callback to be executed
on success/failure. Could we add callback to Sender API? Or should not Aerogear
rather have something like Pipes abstraction for Java developers instead of
pretty dumb Sender API?
Thoughts?
Thanks,
Karel
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
11:59 a.m.
On Tue, 2013-09-17 at 11:22 -0400, Lucas Holmquist wrote:
There server does send a response, "submitted" or something
like that. So maybe there should be callbacks
the node.js sender has callbacks for success and error
+ 1 for callbacks
On Sep 17, 2013, at 11:17 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
> Hi,
>
> I went once again through
> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
> says that Sender API should be fire&forget. It feels more like "maybe
> fire"&forget, for instance it does not say that your credentials were wrong
- or
> it says, you need parse logs to get that information.
>
> If I think about Android, iOS, JS solutions to communicate with
> UnifiedPush we provide - Pipes - they always provide a callback to be executed
> on success/failure. Could we add callback to Sender API? Or should not Aerogear
> rather have something like Pipes abstraction for Java developers instead of
> pretty dumb Sender API?
>
> Thoughts?
>
> Thanks,
>
> Karel
> _______________________________________________
> 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
10:23 a.m.
Callback do not make to much sense for the java implementation IMO (maybe
for the Node implementation it makes more sense) but for the Java
Implementation of the Sender we could discuss on :
- returning the http status code rather than returning void
- have a better exception handling and introduce a Sender specific Exception
On Tue, Sep 17, 2013 at 5:17 PM, Karel Piwko <kpiwko(a)redhat.com> wrote:
Hi,
I went once again through
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
which
says that Sender API should be fire&forget. It feels more like "maybe
fire"&forget, for instance it does not say that your credentials were
wrong - or
it says, you need parse logs to get that information.
If I think about Android, iOS, JS solutions to communicate with
UnifiedPush we provide - Pipes - they always provide a callback to be
executed
on success/failure. Could we add callback to Sender API? Or should not
Aerogear
rather have something like Pipes abstraction for Java developers instead of
pretty dumb Sender API?
Thoughts?
Thanks,
Karel
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
10:46 a.m.
On Tue, 17 Sep 2013 17:23:25 +0200
Sebastien Blanc <scm.blanc(a)gmail.com> wrote:
Callback do not make to much sense for the java implementation IMO
(maybe
for the Node implementation it makes more sense) but for the Java
Implementation of the Sender we could discuss on :
I think callback makes sense even if its synchronous, especially if Sender API
is synchronous as well.
- returning the http status code rather than returning void
How useful is returning a status response without response body?
- have a better exception handling and introduce a Sender specific Exception
+1
On Tue, Sep 17, 2013 at 5:17 PM, Karel Piwko <kpiwko(a)redhat.com> wrote:
> Hi,
>
> I went once again through
> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
> which
> says that Sender API should be fire&forget. It feels more like "maybe
> fire"&forget, for instance it does not say that your credentials were
> wrong - or
> it says, you need parse logs to get that information.
>
> If I think about Android, iOS, JS solutions to communicate with
> UnifiedPush we provide - Pipes - they always provide a callback to be
> executed
> on success/failure. Could we add callback to Sender API? Or should not
> Aerogear
> rather have something like Pipes abstraction for Java developers instead of
> pretty dumb Sender API?
>
> Thoughts?
>
> Thanks,
>
> Karel
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
10:18 p.m.
On Tue, Sep 17, 2013 at 05:23:25PM +0200, Sebastien Blanc wrote:
Callback do not make to much sense for the java implementation IMO
(maybe
for the Node implementation it makes more sense) but for the Java
Implementation of the Sender we could discuss on :
- returning the http status code rather than returning void
-1, Abstraction leak -
If I need to care about status codes, it's better
for me to just go with plain REST calls.
- have a better exception handling and introduce a Sender specific
Exception
One plausible option, but having symmetry over the several clients
isn't
a bad idea, so +1 to callbacks.
--
qmx
10:39 a.m.
On Tue, Sep 17, 2013 at 5:17 PM, Karel Piwko <kpiwko(a)redhat.com> wrote:
Hi,
I went once again through
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
which
says that Sender API should be fire&forget. It feels more like "maybe
fire"&forget, for instance it does not say that your credentials were
wrong
the server returns 401 if credentials are wrong, so that piece does tell
you what's going wrong :-)
- or
it says, you need parse logs to get that information.
If I think about Android, iOS, JS solutions to communicate with
UnifiedPush we provide - Pipes - they always provide a callback to be
executed
on success/failure. Could we add callback to Sender API? Or should not
Aerogear
rather have something like Pipes abstraction for Java developers instead of
pretty dumb Sender API?
we could add some sort of callback:
success --> 200
error 40x/500 status codes
@200: it does not mean too much, the actual processing starts AFTER the
response returns; but it would be a good indicator: yes, I could at least
submit the job to the UnifiedPush server. Currently, I agree, it's more
like.... well did UPS receive my request >
Thoughts?
Thanks,
Karel
_______________________________________________
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
8:31 a.m.
On 09/17/2013 11:17 AM, Karel Piwko wrote:
Hi,
I went once again through
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
says that Sender API should be fire&forget. It feels more like "maybe
fire"&forget, for instance it does not say that your credentials were wrong -
or
it says, you need parse logs to get that information.
If I think about Android, iOS, JS solutions to communicate with
UnifiedPush we provide - Pipes - they always provide a callback to be executed
on success/failure. Could we add callback to Sender API? Or should not Aerogear
rather have something like Pipes abstraction for Java developers instead of
pretty dumb Sender API?
Thoughts?
In a bit of crazy land perhaps the client could keep a web socket or BSD
Socket open to the server which would let it get callbacks about things
that happen further down the tree.
Thanks,
Karel
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
8:37 a.m.
On Wed, 18 Sep 2013 09:31:36 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
On 09/17/2013 11:17 AM, Karel Piwko wrote:
> Hi,
>
> I went once again through
> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
> says that Sender API should be fire&forget. It feels more like "maybe
> fire"&forget, for instance it does not say that your credentials were
wrong
> - or it says, you need parse logs to get that information.
>
> If I think about Android, iOS, JS solutions to communicate with
> UnifiedPush we provide - Pipes - they always provide a callback to be
> executed on success/failure. Could we add callback to Sender API? Or should
> not Aerogear rather have something like Pipes abstraction for Java
> developers instead of pretty dumb Sender API?
>
> Thoughts?
In a bit of crazy land perhaps the client could keep a web socket or BSD
Socket open to the server which would let it get callbacks about things
that happen further down the tree.
Isn't this land called vert.x?
>
> Thanks,
>
> Karel
> _______________________________________________
> 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
9:13 a.m.
On 09/18/2013 09:37 AM, Karel Piwko wrote:
On Wed, 18 Sep 2013 09:31:36 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
> On 09/17/2013 11:17 AM, Karel Piwko wrote:
>> Hi,
>>
>> I went once again through
>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html - which
>> says that Sender API should be fire&forget. It feels more like "maybe
>> fire"&forget, for instance it does not say that your credentials were
wrong
>> - or it says, you need parse logs to get that information.
>>
>> If I think about Android, iOS, JS solutions to communicate with
>> UnifiedPush we provide - Pipes - they always provide a callback to be
>> executed on success/failure. Could we add callback to Sender API? Or should
>> not Aerogear rather have something like Pipes abstraction for Java
>> developers instead of pretty dumb Sender API?
>>
>> Thoughts?
> In a bit of crazy land perhaps the client could keep a web socket or BSD
> Socket open to the server which would let it get callbacks about things
> that happen further down the tree.
Isn't this land called vert.x?
Maybe I misunderstood. I thought it was wanting
to get information from
the push server about the status of messages being sent not the response
of the commands to the push server itself.
>> Thanks,
>>
>> Karel
>> _______________________________________________
>> 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
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
9:16 a.m.
On Thu, Sep 19, 2013 at 4:13 PM, Summers Pittman <supittma(a)redhat.com>wrote:
On 09/18/2013 09:37 AM, Karel Piwko wrote:
> On Wed, 18 Sep 2013 09:31:36 -0400
> Summers Pittman <supittma(a)redhat.com> wrote:
>
>> On 09/17/2013 11:17 AM, Karel Piwko wrote:
>>> Hi,
>>>
>>> I went once again through
>>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
which
>>> says that Sender API should be fire&forget. It feels more like
"maybe
>>> fire"&forget, for instance it does not say that your credentials
were
wrong
>>> - or it says, you need parse logs to get that information.
>>>
>>> If I think about Android, iOS, JS solutions to communicate with
>>> UnifiedPush we provide - Pipes - they always provide a callback to be
>>> executed on success/failure. Could we add callback to Sender API? Or
should
>>> not Aerogear rather have something like Pipes abstraction for Java
>>> developers instead of pretty dumb Sender API?
>>>
>>> Thoughts?
>> In a bit of crazy land perhaps the client could keep a web socket or BSD
>> Socket open to the server which would let it get callbacks about things
>> that happen further down the tree.
> Isn't this land called vert.x?
Maybe I misunderstood. I thought it was wanting to get information from
the push server about the status of messages being sent not the response
of the commands to the push server itself.
well, all the server tells is:
* job submit for processing (aka I am now dispatching to Google, Apple etc)
* 401 auth issue
* 500 errors
-M
>
>>> Thanks,
>>>
>>> Karel
>>> _______________________________________________
>>> 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
> _______________________________________________
> 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
4:43 a.m.
On Thu, 19 Sep 2013 10:13:52 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
On 09/18/2013 09:37 AM, Karel Piwko wrote:
> On Wed, 18 Sep 2013 09:31:36 -0400
> Summers Pittman <supittma(a)redhat.com> wrote:
>
>> On 09/17/2013 11:17 AM, Karel Piwko wrote:
>>> Hi,
>>>
>>> I went once again through
>>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
>>> which says that Sender API should be fire&forget. It feels more like
>>> "maybe fire"&forget, for instance it does not say that your
credentials
>>> were wrong
>>> - or it says, you need parse logs to get that information.
>>>
>>> If I think about Android, iOS, JS solutions to communicate with
>>> UnifiedPush we provide - Pipes - they always provide a callback to be
>>> executed on success/failure. Could we add callback to Sender API? Or
>>> should not Aerogear rather have something like Pipes abstraction for Java
>>> developers instead of pretty dumb Sender API?
>>>
>>> Thoughts?
>> In a bit of crazy land perhaps the client could keep a web socket or BSD
>> Socket open to the server which would let it get callbacks about things
>> that happen further down the tree.
> Isn't this land called vert.x?
Maybe I misunderstood. I thought it was wanting to get information from
the push server about the status of messages being sent not the response
of the commands to the push server itself.
I was speaking about the latter.
>
>>> Thanks,
>>>
>>> Karel
>>> _______________________________________________
>>> 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
> _______________________________________________
> 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
8:14 a.m.
FYI Karel has created a Jira for this
https://issues.jboss.org/browse/AGPUSH-373
On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
On Thu, 19 Sep 2013 10:13:52 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
> On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > On Wed, 18 Sep 2013 09:31:36 -0400
> > Summers Pittman <supittma(a)redhat.com> wrote:
> >
> >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> >>> Hi,
> >>>
> >>> I went once again through
> >>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html-
> >>> which says that Sender API should be fire&forget. It feels more
like
> >>> "maybe fire"&forget, for instance it does not say that
your
credentials
> >>> were wrong
> >>> - or it says, you need parse logs to get that information.
> >>>
> >>> If I think about Android, iOS, JS solutions to communicate with
> >>> UnifiedPush we provide - Pipes - they always provide a callback to be
> >>> executed on success/failure. Could we add callback to Sender API? Or
> >>> should not Aerogear rather have something like Pipes abstraction for
Java
> >>> developers instead of pretty dumb Sender API?
> >>>
> >>> Thoughts?
> >> In a bit of crazy land perhaps the client could keep a web socket or
BSD
> >> Socket open to the server which would let it get callbacks about
things
> >> that happen further down the tree.
> > Isn't this land called vert.x?
> Maybe I misunderstood. I thought it was wanting to get information from
> the push server about the status of messages being sent not the response
> of the commands to the push server itself.
I was speaking about the latter.
> >
> >>> Thanks,
> >>>
> >>> Karel
> >>> _______________________________________________
> >>> 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
> > _______________________________________________
> > 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
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
6:21 a.m.
So I started to look more closely to this, and I'm wondering how we want
the Callback to be. I see 2 options that are described here :
https://gist.github.com/sebastienblanc/7023151
Basically, version 1 has a 'completed' method which will be invoked no
matter which http status code is returned, the developer has than to
implement it's own logic of handling the status. The 'failure' method will
be invoked in case of an exception is thrown (IOException for instance)
Version 2 has a 'onSuccess' method which handles the 2xx response codes and
a 'onError' which handles 4xx, 5xx codes or even if an exception
has occurred.
Do you have any preference or even an alternative solution ?
Seb
On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com> wrote:
FYI Karel has created a Jira for this
https://issues.jboss.org/browse/AGPUSH-373
On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
> On Thu, 19 Sep 2013 10:13:52 -0400
> Summers Pittman <supittma(a)redhat.com> wrote:
>
> > On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > > On Wed, 18 Sep 2013 09:31:36 -0400
> > > Summers Pittman <supittma(a)redhat.com> wrote:
> > >
> > >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> > >>> Hi,
> > >>>
> > >>> I went once again through
> > >>>
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html-
> > >>> which says that Sender API should be fire&forget. It feels more
like
> > >>> "maybe fire"&forget, for instance it does not say
that your
> credentials
> > >>> were wrong
> > >>> - or it says, you need parse logs to get that information.
> > >>>
> > >>> If I think about Android, iOS, JS solutions to communicate with
> > >>> UnifiedPush we provide - Pipes - they always provide a callback to
> be
> > >>> executed on success/failure. Could we add callback to Sender API?
Or
> > >>> should not Aerogear rather have something like Pipes abstraction
> for Java
> > >>> developers instead of pretty dumb Sender API?
> > >>>
> > >>> Thoughts?
> > >> In a bit of crazy land perhaps the client could keep a web socket or
> BSD
> > >> Socket open to the server which would let it get callbacks about
> things
> > >> that happen further down the tree.
> > > Isn't this land called vert.x?
> > Maybe I misunderstood. I thought it was wanting to get information from
> > the push server about the status of messages being sent not the response
> > of the commands to the push server itself.
>
> I was speaking about the latter.
>
> > >
> > >>> Thanks,
> > >>>
> > >>> Karel
> > >>> _______________________________________________
> > >>> 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
> > > _______________________________________________
> > > 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
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
6:25 a.m.
+ 1 on letting the client decide what status is considered as success
(version 1)
On Thu, 2013-10-17 at 13:21 +0200, Sebastien Blanc wrote:
So I started to look more closely to this, and I'm wondering how
we
want the Callback to be. I see 2 options that are described
here : https://gist.github.com/sebastienblanc/7023151
Basically, version 1 has a 'completed' method which will be invoked no
matter which http status code is returned, the developer has than to
implement it's own logic of handling the status. The 'failure' method
will be invoked in case of an exception is thrown (IOException for
instance)
Version 2 has a 'onSuccess' method which handles the 2xx response
codes and a 'onError' which handles 4xx, 5xx codes or even if an
exception has occurred.
Do you have any preference or even an alternative solution ?
Seb
On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com>
wrote:
FYI Karel has created a Jira for
this https://issues.jboss.org/browse/AGPUSH-373
On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko
<kpiwko(a)redhat.com> wrote:
On Thu, 19 Sep 2013 10:13:52 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
> On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > On Wed, 18 Sep 2013 09:31:36 -0400
> > Summers Pittman <supittma(a)redhat.com> wrote:
> >
> >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> >>> Hi,
> >>>
> >>> I went once again through
> >>>
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
> >>> which says that Sender API should be
fire&forget. It feels more like
> >>> "maybe fire"&forget, for instance it does
not
say that your credentials
> >>> were wrong
> >>> - or it says, you need parse logs to get that
information.
> >>>
> >>> If I think about Android, iOS, JS solutions to
communicate with
> >>> UnifiedPush we provide - Pipes - they always
provide a callback to be
> >>> executed on success/failure. Could we add
callback to Sender API? Or
> >>> should not Aerogear rather have something like
Pipes abstraction for Java
> >>> developers instead of pretty dumb Sender API?
> >>>
> >>> Thoughts?
> >> In a bit of crazy land perhaps the client could
keep a web socket or BSD
> >> Socket open to the server which would let it get
callbacks about things
> >> that happen further down the tree.
> > Isn't this land called vert.x?
> Maybe I misunderstood. I thought it was wanting to
get information from
> the push server about the status of messages being
sent not the response
> of the commands to the push server itself.
I was speaking about the latter.
> >
> >>> Thanks,
> >>>
> >>> Karel
> >>> _______________________________________________
> >>> 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
> > _______________________________________________
> > 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
_______________________________________________
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
7:20 a.m.
I'm a bit partial to option 2 since that is how we do it in JS, however; there is a
pattern in node where there is one callback that is passed an error argument( if available
). something like this gist
https://gist.github.com/lholmquist/7023904
although either way, we are determining what is an error
On Oct 17, 2013, at 7:25 AM, Apostolos Emmanouilidis <aemmanou(a)redhat.com> wrote:
+ 1 on letting the client decide what status is considered as success
(version 1)
On Thu, 2013-10-17 at 13:21 +0200, Sebastien Blanc wrote:
>
> So I started to look more closely to this, and I'm wondering how we want the
Callback to be. I see 2 options that are described here :
https://gist.github.com/sebastienblanc/7023151
>
>
> Basically, version 1 has a 'completed' method which will be invoked no matter
which http status code is returned, the developer has than to implement it's own logic
of handling the status. The 'failure' method will be invoked in case of an
exception is thrown (IOException for instance)
>
>
> Version 2 has a 'onSuccess' method which handles the 2xx response codes and a
'onError' which handles 4xx, 5xx codes or even if an exception has occurred.
>
>
> Do you have any preference or even an alternative solution ?
>
>
> Seb
>
>
>
>
>
>
>
>
>
> On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com> wrote:
> FYI Karel has created a Jira for this https://issues.jboss.org/browse/AGPUSH-373
>
>
>
>
> On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
> On Thu, 19 Sep 2013 10:13:52 -0400
> Summers Pittman <supittma(a)redhat.com> wrote:
>
> > On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > > On Wed, 18 Sep 2013 09:31:36 -0400
> > > Summers Pittman <supittma(a)redhat.com> wrote:
> > >
> > >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> > >>> Hi,
> > >>>
> > >>> I went once again through
> > >>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html
-
> > >>> which says that Sender API should be fire&forget. It feels more
like
> > >>> "maybe fire"&forget, for instance it does not say
that your credentials
> > >>> were wrong
> > >>> - or it says, you need parse logs to get that information.
> > >>>
> > >>> If I think about Android, iOS, JS solutions to communicate with
> > >>> UnifiedPush we provide - Pipes - they always provide a callback to
be
> > >>> executed on success/failure. Could we add callback to Sender API?
Or
> > >>> should not Aerogear rather have something like Pipes abstraction
for Java
> > >>> developers instead of pretty dumb Sender API?
> > >>>
> > >>> Thoughts?
> > >> In a bit of crazy land perhaps the client could keep a web socket or
BSD
> > >> Socket open to the server which would let it get callbacks about
things
> > >> that happen further down the tree.
> > > Isn't this land called vert.x?
> > Maybe I misunderstood. I thought it was wanting to get information from
> > the push server about the status of messages being sent not the response
> > of the commands to the push server itself.
>
>
> I was speaking about the latter.
>
> > >
> > >>> Thanks,
> > >>>
> > >>> Karel
> > >>> _______________________________________________
> > >>> 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
> > > _______________________________________________
> > > 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
>
> _______________________________________________
> 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
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
7:23 a.m.
We could also propose both in the same callback, a bit ala JavaScript :
meaning offering also an 'onSuccess' (for 2xx) with still the 'complete'
which will be called anyway. Could be a nice convenience method for
slackers.
On Thu, Oct 17, 2013 at 2:20 PM, Lucas Holmquist <lholmqui(a)redhat.com>wrote:
I'm a bit partial to option 2 since that is how we do it in JS,
however;
there is a pattern in node where there is one callback that is passed an
error argument( if available ). something like this gist
https://gist.github.com/lholmquist/7023904
although either way, we are determining what is an error
On Oct 17, 2013, at 7:25 AM, Apostolos Emmanouilidis <aemmanou(a)redhat.com>
wrote:
+ 1 on letting the client decide what status is considered as success
(version 1)
On Thu, 2013-10-17 at 13:21 +0200, Sebastien Blanc wrote:
So I started to look more closely to this, and I'm wondering how we want
the Callback to be. I see 2 options that are described here :
https://gist.github.com/sebastienblanc/7023151
Basically, version 1 has a 'completed' method which will be invoked no
matter which http status code is returned, the developer has than to
implement it's own logic of handling the status. The 'failure' method will
be invoked in case of an exception is thrown (IOException for instance)
Version 2 has a 'onSuccess' method which handles the 2xx response codes
and a 'onError' which handles 4xx, 5xx codes or even if an exception
has occurred.
Do you have any preference or even an alternative solution ?
Seb
On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com>
wrote:
FYI Karel has created a Jira for this
https://issues.jboss.org/browse/AGPUSH-373
On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
On Thu, 19 Sep 2013 10:13:52 -0400
Summers Pittman <supittma(a)redhat.com> wrote:
> On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > On Wed, 18 Sep 2013 09:31:36 -0400
> > Summers Pittman <supittma(a)redhat.com> wrote:
> >
> >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> >>> Hi,
> >>>
> >>> I went once again through
> >>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html-
> >>> which says that Sender API should be fire&forget. It feels more
like
> >>> "maybe fire"&forget, for instance it does not say that
your
credentials
> >>> were wrong
> >>> - or it says, you need parse logs to get that information.
> >>>
> >>> If I think about Android, iOS, JS solutions to communicate with
> >>> UnifiedPush we provide - Pipes - they always provide a callback to be
> >>> executed on success/failure. Could we add callback to Sender API? Or
> >>> should not Aerogear rather have something like Pipes abstraction for
Java
> >>> developers instead of pretty dumb Sender API?
> >>>
> >>> Thoughts?
> >> In a bit of crazy land perhaps the client could keep a web socket or
BSD
> >> Socket open to the server which would let it get callbacks about
things
> >> that happen further down the tree.
> > Isn't this land called vert.x?
> Maybe I misunderstood. I thought it was wanting to get information from
> the push server about the status of messages being sent not the response
> of the commands to the push server itself.
I was speaking about the latter.
> >
> >>> Thanks,
> >>>
> >>> Karel
> >>> _______________________________________________
> >>> 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
> > _______________________________________________
> > 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
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
_______________________________________________
aerogear-dev mailing
listaerogear-dev@lists.jboss.orghttps://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
_______________________________________________
aerogear-dev mailing list
aerogear-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/aerogear-dev
8:14 a.m.
To give a bit more concrete "material" , I've submitted a first PR
https://github.com/aerogear/aerogear-unifiedpush-java-client/pull/24 ,
please feel free to comment on it on specific points, for more general
discussions let's keep using this thread.
On Thu, Oct 17, 2013 at 2:23 PM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
We could also propose both in the same callback, a bit ala JavaScript
:
meaning offering also an 'onSuccess' (for 2xx) with still the 'complete'
which will be called anyway. Could be a nice convenience method for
slackers.
On Thu, Oct 17, 2013 at 2:20 PM, Lucas Holmquist <lholmqui(a)redhat.com>wrote:
> I'm a bit partial to option 2 since that is how we do it in JS, however;
> there is a pattern in node where there is one callback that is passed an
> error argument( if available ). something like this gist
> https://gist.github.com/lholmquist/7023904
>
> although either way, we are determining what is an error
>
> On Oct 17, 2013, at 7:25 AM, Apostolos Emmanouilidis <aemmanou(a)redhat.com>
> wrote:
>
> + 1 on letting the client decide what status is considered as success
> (version 1)
>
> On Thu, 2013-10-17 at 13:21 +0200, Sebastien Blanc wrote:
>
> So I started to look more closely to this, and I'm wondering how we want
> the Callback to be. I see 2 options that are described here :
> https://gist.github.com/sebastienblanc/7023151
>
>
>
> Basically, version 1 has a 'completed' method which will be invoked no
> matter which http status code is returned, the developer has than to
> implement it's own logic of handling the status. The 'failure' method
will
> be invoked in case of an exception is thrown (IOException for instance)
>
>
>
> Version 2 has a 'onSuccess' method which handles the 2xx response codes
> and a 'onError' which handles 4xx, 5xx codes or even if an exception
> has occurred.
>
>
>
> Do you have any preference or even an alternative solution ?
>
>
>
> Seb
>
>
>
>
>
>
>
>
>
>
> On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com>
> wrote:
>
> FYI Karel has created a Jira for this
> https://issues.jboss.org/browse/AGPUSH-373
>
>
>
>
>
> On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com>
> wrote:
>
> On Thu, 19 Sep 2013 10:13:52 -0400
>
> Summers Pittman <supittma(a)redhat.com> wrote:
>
> > On 09/18/2013 09:37 AM, Karel Piwko wrote:
> > > On Wed, 18 Sep 2013 09:31:36 -0400
> > > Summers Pittman <supittma(a)redhat.com> wrote:
> > >
> > >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
> > >>> Hi,
> > >>>
> > >>> I went once again through
> > >>>
http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html-
> > >>> which says that Sender API should be fire&forget. It feels more
like
> > >>> "maybe fire"&forget, for instance it does not say
that your
> credentials
> > >>> were wrong
> > >>> - or it says, you need parse logs to get that information.
> > >>>
> > >>> If I think about Android, iOS, JS solutions to communicate with
> > >>> UnifiedPush we provide - Pipes - they always provide a callback to
> be
> > >>> executed on success/failure. Could we add callback to Sender API?
Or
> > >>> should not Aerogear rather have something like Pipes abstraction
> for Java
> > >>> developers instead of pretty dumb Sender API?
> > >>>
> > >>> Thoughts?
> > >> In a bit of crazy land perhaps the client could keep a web socket or
> BSD
> > >> Socket open to the server which would let it get callbacks about
> things
> > >> that happen further down the tree.
> > > Isn't this land called vert.x?
> > Maybe I misunderstood. I thought it was wanting to get information from
> > the push server about the status of messages being sent not the response
> > of the commands to the push server itself.
>
>
> I was speaking about the latter.
>
>
> > >
> > >>> Thanks,
> > >>>
> > >>> Karel
> > >>> _______________________________________________
> > >>> 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
> > > _______________________________________________
> > > 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
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
>
>
>
>
> _______________________________________________
> aerogear-dev mailing
listaerogear-dev@lists.jboss.orghttps://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
>
>
>
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev
>
7:19 a.m.
On Thu, Oct 17, 2013 at 1:21 PM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
So I started to look more closely to this, and I'm wondering how
we want
the Callback to be. I see 2 options that are described here :
https://gist.github.com/sebastienblanc/7023151
Basically, version 1 has a 'completed' method which will be invoked no
matter which http status code is returned, the developer has than to
implement it's own logic of handling the status. The 'failure' method will
be invoked in case of an exception is thrown (IOException for instance)
Similar to what the JAX-RS 2.0 client lib does:
https://jersey.java.net/documentation/latest/async.html#d0e7394
Will this 'dramatically' change our call as well (e.g. async nature?)
-M
Version 2 has a 'onSuccess' method which handles the 2xx response codes
and a 'onError' which handles 4xx, 5xx codes or even if an exception
has occurred.
Do you have any preference or even an alternative solution ?
Seb
On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
> FYI Karel has created a Jira for this
> https://issues.jboss.org/browse/AGPUSH-373
>
>
>
> On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
>
>> On Thu, 19 Sep 2013 10:13:52 -0400
>> Summers Pittman <supittma(a)redhat.com> wrote:
>>
>> > On 09/18/2013 09:37 AM, Karel Piwko wrote:
>> > > On Wed, 18 Sep 2013 09:31:36 -0400
>> > > Summers Pittman <supittma(a)redhat.com> wrote:
>> > >
>> > >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
>> > >>> Hi,
>> > >>>
>> > >>> I went once again through
>> > >>>
>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
>> > >>> which says that Sender API should be fire&forget. It feels
more
>> like
>> > >>> "maybe fire"&forget, for instance it does not say
that your
>> credentials
>> > >>> were wrong
>> > >>> - or it says, you need parse logs to get that information.
>> > >>>
>> > >>> If I think about Android, iOS, JS solutions to communicate
with
>> > >>> UnifiedPush we provide - Pipes - they always provide a callback
to
>> be
>> > >>> executed on success/failure. Could we add callback to Sender
API?
>> Or
>> > >>> should not Aerogear rather have something like Pipes
abstraction
>> for Java
>> > >>> developers instead of pretty dumb Sender API?
>> > >>>
>> > >>> Thoughts?
>> > >> In a bit of crazy land perhaps the client could keep a web socket
>> or BSD
>> > >> Socket open to the server which would let it get callbacks about
>> things
>> > >> that happen further down the tree.
>> > > Isn't this land called vert.x?
>> > Maybe I misunderstood. I thought it was wanting to get information
>> from
>> > the push server about the status of messages being sent not the
>> response
>> > of the commands to the push server itself.
>>
>> I was speaking about the latter.
>>
>> > >
>> > >>> Thanks,
>> > >>>
>> > >>> Karel
>> > >>> _______________________________________________
>> > >>> 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
>> > > _______________________________________________
>> > > 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
>>
>> _______________________________________________
>> 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
7:22 a.m.
On Thu, Oct 17, 2013 at 2:19 PM, Matthias Wessendorf <matzew(a)apache.org>wrote:
On Thu, Oct 17, 2013 at 1:21 PM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
> So I started to look more closely to this, and I'm wondering how we want
> the Callback to be. I see 2 options that are described here :
> https://gist.github.com/sebastienblanc/7023151
>
> Basically, version 1 has a 'completed' method which will be invoked no
> matter which http status code is returned, the developer has than to
> implement it's own logic of handling the status. The 'failure' method
will
> be invoked in case of an exception is thrown (IOException for instance)
>
Similar to what the JAX-RS 2.0 client lib does:
https://jersey.java.net/documentation/latest/async.html#d0e7394
Will this 'dramatically' change our call as well (e.g. async nature?)
Nope, both options have no big impact.
But speaking of that, I plan to keep the old merthod send(UnifiedMessage
message), the dev can choose just a fire&forget or with callback ? wdyt ?
-M
>
> Version 2 has a 'onSuccess' method which handles the 2xx response codes
> and a 'onError' which handles 4xx, 5xx codes or even if an exception
> has occurred.
>
> Do you have any preference or even an alternative solution ?
>
> Seb
>
>
>
>
>
>
> On Tue, Oct 8, 2013 at 3:14 PM, Sebastien Blanc <scm.blanc(a)gmail.com>wrote:
>
>> FYI Karel has created a Jira for this
>> https://issues.jboss.org/browse/AGPUSH-373
>>
>>
>>
>> On Tue, Sep 24, 2013 at 11:43 AM, Karel Piwko <kpiwko(a)redhat.com> wrote:
>>
>>> On Thu, 19 Sep 2013 10:13:52 -0400
>>> Summers Pittman <supittma(a)redhat.com> wrote:
>>>
>>> > On 09/18/2013 09:37 AM, Karel Piwko wrote:
>>> > > On Wed, 18 Sep 2013 09:31:36 -0400
>>> > > Summers Pittman <supittma(a)redhat.com> wrote:
>>> > >
>>> > >> On 09/17/2013 11:17 AM, Karel Piwko wrote:
>>> > >>> Hi,
>>> > >>>
>>> > >>> I went once again through
>>> > >>>
>>> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
>>> > >>> which says that Sender API should be fire&forget. It
feels more
>>> like
>>> > >>> "maybe fire"&forget, for instance it does not
say that your
>>> credentials
>>> > >>> were wrong
>>> > >>> - or it says, you need parse logs to get that information.
>>> > >>>
>>> > >>> If I think about Android, iOS, JS solutions to communicate
with
>>> > >>> UnifiedPush we provide - Pipes - they always provide a
callback
>>> to be
>>> > >>> executed on success/failure. Could we add callback to
Sender API?
>>> Or
>>> > >>> should not Aerogear rather have something like Pipes
abstraction
>>> for Java
>>> > >>> developers instead of pretty dumb Sender API?
>>> > >>>
>>> > >>> Thoughts?
>>> > >> In a bit of crazy land perhaps the client could keep a web
socket
>>> or BSD
>>> > >> Socket open to the server which would let it get callbacks
about
>>> things
>>> > >> that happen further down the tree.
>>> > > Isn't this land called vert.x?
>>> > Maybe I misunderstood. I thought it was wanting to get information
>>> from
>>> > the push server about the status of messages being sent not the
>>> response
>>> > of the commands to the push server itself.
>>>
>>> I was speaking about the latter.
>>>
>>> > >
>>> > >>> Thanks,
>>> > >>>
>>> > >>> Karel
>>> > >>> _______________________________________________
>>> > >>> 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
>>> > > _______________________________________________
>>> > > 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
>>>
>>> _______________________________________________
>>> 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
8:37 a.m.
On Wed, Sep 18, 2013 at 3:31 PM, Summers Pittman <supittma(a)redhat.com>wrote:
On 09/17/2013 11:17 AM, Karel Piwko wrote:
> Hi,
>
> I went once again through
> http://lists.jboss.org/pipermail/aerogear-dev/2013-June/002901.html -
which
> says that Sender API should be fire&forget. It feels more like "maybe
> fire"&forget, for instance it does not say that your credentials were
wrong - or
> it says, you need parse logs to get that information.
>
> If I think about Android, iOS, JS solutions to communicate with
> UnifiedPush we provide - Pipes - they always provide a callback to be
executed
> on success/failure. Could we add callback to Sender API? Or should not
Aerogear
> rather have something like Pipes abstraction for Java developers instead
of
> pretty dumb Sender API?
>
> Thoughts?
In a bit of crazy land perhaps the client could keep a web socket or BSD
Socket open to the server which would let it get callbacks about things
that happen further down the tree.
-1
>
> Thanks,
>
> Karel
> _______________________________________________
> 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
4054
days inactive
4085
days old
20 comments
7 participants
participants (7)
-
Apostolos Emmanouilidis -
Douglas Campos -
Karel Piwko -
Lucas Holmquist -
Matthias Wessendorf -
Sebastien Blanc -
Summers Pittman