Deepal, let's review the status early morning on Friday. My biggest
concern now is the security stuff (authn) and getting it done right too.
Sanjiva.
Deepal Jayasinghe wrote:
I have completed most of the APP stuff according to the wiki and will
commit the complete code soon , so that you can comment on that and I
can continue the rest.
-Deepal
I went though the code and looked what actually I have done then I found
that the only problems in the resource operations like raring and
tagging. Adding , delete and accessing resources does not have any
issues, and it has follow the API in the wiki correctly [1] . Now I
started to change the implementation to cope with the API in the wiki ,
but I have a number of area which I have doubt and need to clear them
before I continue.
- How to get logs
should that be /[r1];logs
- Get versions
/[r1];versions , then the text of the feed will be the versions for
the given resource.
- /[r1];tags:[tag+username] , I can not understand the use of this
-For the rating and tagging what should be the structure of the entry .
- Did we finalized on ";" as the parameter separator.
[1] : http://wso2.org/wiki/display/registry/Registry+Protocol
-Thanks
Deepal
It is essential that we have a *perfect* implementation of APP here
and that it is clearly documented and /matches the documentation/!
We have deliberately avoided the UDDI crowd here... so we can't afford
to annoy the REST community by bodging APP :)
Paul
Glen Daniels wrote:
Hi Deepal, all:
Deepal Jayasinghe wrote:
Are we not using Abdera on the client side?
We do.
Yep - and upon further research into the code it looks like we
haven't paid much attention to the protocol design we did earlier:
http://wso2.org/wiki/display/registry/Registry+Protocol
Not only do we have the non-APP-ish use of POST to non-existent URLs
(in order to create them), but for instance to tag, we seem to do a
PUT of an Atom entry representing the tag to the resource URL -
shouldn't that be a POST to "...resource;tags"?
(Looks like we are given the user agent header.) So doesn't Abdera do
the right thing for this??
Abdera is apparently a little (too?) flexible about this kind of stuff.
Also, I noticed that in the code below we connect to the registry at
one URL but the base URL for the Atom stuff has "/atom" added to the
reg URL passed in. Is that right? Should we not say the base URL is
.../wso2registry/atom instead?
We can do that . But I intentionally implement the code to give the URL
of the registry not the URL of the ATOM. Because user does not want to
know whether we use APP or not , he just need a remote API to talk to a
registry. So I personally do not like to provide /atom when we give
the
url.
+1 to not providing /atom when we give the URL. But -1 to tacking it
on in the first place. There is no need for the registry API to make
any assumptions about the URL except that it's rooted wherever we're
told. In other words - shouldn't I be able to say:
Registry R1 = new RemoteRegistry("http://myhost/regRoot";);
Registry R2 = new RemoteRegistry("http://myhost/regRoot/subDir";);
...and have both R1 and R2 work? Isn't this how we're expecting to
use this for things like Synapse/Axis2 repositories?
new RemoteRegistry("http://registrySite/registry/finance/axis2repo";);
Thoughts?
--Glen
_______________________________________________
Registry-dev mailing list
[email protected]
http://wso2.org/cgi-bin/mailman/listinfo/registry-dev
_______________________________________________
Registry-dev mailing list
[email protected]
http://wso2.org/cgi-bin/mailman/listinfo/registry-dev
--
Sanjiva Weerawarana, Ph.D.
Founder, Chairman & CEO; WSO2, Inc.; http://www.wso2.com/
email: [EMAIL PROTECTED]; cell: +1 650 265 8311 | +94 77 787 6880
"Oxygenating the Web Service Platform."
_______________________________________________
Registry-dev mailing list
[email protected]
http://wso2.org/cgi-bin/mailman/listinfo/registry-dev
