Oauth question...

Someone was saying fetchmail would need to do a client request.

Presumably that sends back a response asking the oauth provider if the 
client
should have access.

Then that sends back a 'blob' (cookie) that is used for future access?

Also was something about the need for a json parser.

Is that it?

The display of the response seems to be 1 sticking point -- wouldn't sending
the response to a local system or user defined browser? to handle the Q+A,
then the stub stores the response for future authorizations?

Or how much more complex is oauth than I'm thinking...?

Something that parses limited HTML isn't real hard in perl.  I have a 
few local apps that do such parsing,
like to pull down rss_rdr, or one to check for the latest linux kernel 
version, etc.

The part that might get hairy is invoking a local client to display the 
web-auth page and get the
blob back from that transaction (?)  But isn't the rest mostly straight 
forward?

It may be beyond my capabilities, but I wouldn't know w/o a better idea 
of the problem.

All of that code is mostly for setup.

It seems once you have the token, no more interactive browser access is 
needed, is that
about right?

Sorry if this has been designed in detail for fetchmail already ...if 
so, is there
a pointer to the design?


Thanks,
-linda