2010/8/28 Steve Vinoski <[a:mailto:vinoski@...>
>2010/8/12 Davide Marquês <nesrait@...>:
>> I won't be able to help you sort this out at this point but hopefully other
>> mailing list lurkers can lend an hand and/or step in and update the current
>> implementation. ;)
>Dominique Boucher has stepped up to the plate and updated websockets
>for us. I just pushed the changes to github a few minutes ago. Thanks,
Thank you Dominique.
Thank you Steve, Davide and everyone. :)
It seems much better than my experimental patch.
I tested it on my application, and safri 5.0.1 works fine.
Now, we have no reason to be afraid of mac users and "draft 76". ;)
But in my test,
yaws_api:websocket_unframe_data/1 still fails sometime.
I think this is problem of other part of yaws. ???
And I have an idea about usage of md5.
On my windows environment, crypto:md5_xxx fails because of
an absence of openssl(ssh?).
So I think it better to use erlang:md5/1 like this.
challenge(Key1, Key2, Challenge) ->
% BinaryAnswer =
% digits32(Key1) ++ digits32(Key2) ++ Challenge)),
erlang:md5( digits32(Key1) ++ digits32(Key2) ++ Challenge ).