Open GoogleCodeExporter opened 9 years ago
I'm not sure you can call this a bug... If x participant is not part of a wave,
why should they be able to fetch it?
Simple solution: Add the proxy first, then call submit(wavelet), then call
fetch_wavelet for the proxy.
Original comment by nat.abbo...@gmail.com
on 28 Jul 2010 at 8:37
Yes, I could do that, but I think that the proxy_for parameter in fetch_wavelet
would be more or less useless.
With this method I have to:
1. fetch the wavelet
2. add the participant
3. submit the wavelet
5. fetch it again with the proxy_id
Where I could use the following method:
1. fetch the wavelet
2. add the robot address to the wavelet
3. use wavelet.proxy_for
Or even better:
1. fetch the wavelet with the proxy_id ... and done
In my opinion myRobot.fetch_wavelet("SomeWaveId", None, "proxyid") is just a
shortcut for the last method.
If I first have to fetch the wavelet and add the participant, why should I use
fetch_wavelet with a proxy at all and not just use the method with
wavelet.proxy_for?
I think fetch_wavelet should work the same way as wavelet.proxy_for. Why should
I need to add the proxy participant if I use fetch_wavelet but not if I use
wavelet.proxy_for?
Does anybody know how this works in the Java API?
Original comment by lehrg...@gmail.com
on 28 Jul 2010 at 9:49
Also - this isn't necessary at all - you can actually do what you want already.
Simply add the proxy to the wavelet and then call wavelet.proxyFor('proxyid')
where 'proxyid' only the part between the + & the @ in the following
robot+proxyid@appspot.com
Original comment by nat.abbo...@gmail.com
on 28 Jul 2010 at 1:03
I can even do that _without_ adding the proxy to the wavelet first! And that is
where the fetch_wavelet API has a bug or is at least inconsistent with the rest
of the API.
What I am doing at the moment ist:
1. fetch_wavelet without proxy id
2. add robot address to wavelet
3. call proxy_for on the wavelet with the proxy_id (-> no need to add the proxy
id to the wavelet before that! It just adds it _automatically_.)
What I want is:
1. fetch_wavelet with proxy id (_without_ adding the participant prior to this
_manually_, why else would I use fetch wavelet with the proxy_id?)
Original comment by lehrg...@gmail.com
on 28 Jul 2010 at 1:58
But fetch_wavelet gets you a copy of the wavelet from x-proxy's point of view.
If x-proxy isn't on the wavelet, it has no point of view, so you get an error.
So why can't you just fetch_wavelet w/o proxy, then call proxy for on the
wavelet? What you are trying to do defies the nature of wave IMHO.
Original comment by nat.abbo...@gmail.com
on 28 Jul 2010 at 4:05
Sure I can do it this way, but why has fetch_wavelet a proxy_for param if it is
more or less useless?
At the moment if I ever want to use the fetch_wavelet with a proxy_id a have to
fetch the wavelet first without a proxy id to add the proxy user to it or check
if the proxy user is already a participant of the wavelet.
And if this defies the nature of wave than the method with wavelet.proxy_for
defies it too, because the proxy user is added automatically.
IMO both methods should behave the same and the documentation should be clear
about what the prerequisites of this functions are (for example a already added
participant) or state that the participant will be added automatically.
Original comment by lehrg...@gmail.com
on 28 Jul 2010 at 4:40
Fetch wavelet has a proxy for parameter so that you can fetch a wavelet on
behalf of a proxy that is on that wave.
Original comment by nat.abbo...@gmail.com
on 28 Jul 2010 at 5:35
OK, but how should a programmer know if a proxy_id (just some kind of a string)
is on the wavelet or not? To figure this out a programmer has to fetch the
wavelet and check the list of participants or maintain a list of participant
outside of wave...
I sill think this is a bug. Especially since you do not need to manually add
the proxy participant with wavelet.proxy_for.
How does fetchWavelet and wavelet.proxyFor works in the Java API?
Original comment by lehrg...@gmail.com
on 28 Jul 2010 at 7:36
Original issue reported on code.google.com by
lehrg...@gmail.com
on 19 Jul 2010 at 10:02