Looks good to me! Thanks for making this available.
Two niggly code comments:
1) Why give [session.Identifier] an [option] type? It seems always to
be [Some] here.
2) Near the end of the .ur file appears the code [v <- signal s; return
v]. By the usual monad laws, this is equivalent to [signal s] by itself.
Do you have plans to package this code as a library, encapsulating
protocol details much as the openid library does? I can see it making
sense to expand the openid library to cover a variety of protocols,
possibly also including Facebook Connect. It would also be nice if
someone dove into the current openid C code to fix a longstanding bug
with the OpenID crypto setting that LiveJournal uses but no one else
seems to. ;)
I probably won't work on any of the above myself in the near future, but
I'd be glad to provide advice, accept patches, etc.
On 11/13/2013 02:19 AM, Chris Double wrote:
I've been wanting to use the Persona Authentication API [1] instead of
OpenID. I wrote an example based on the Persona Quick Setup [2] and put
it on github:
<https://github.com/doublec/urweb-persona>
For 'real world' usage it'd probably not be too much work to either add
or modify the openid ur/web code to use persona.
Feedback on the approach I've taken is welcome.
[1] https://login.persona.org/
[2] https://developer.mozilla.org/en-US/Persona/Quick_Setup
_______________________________________________
Ur mailing list
[email protected]
http://www.impredicative.com/cgi-bin/mailman/listinfo/ur
