Autoconfig for aol.com and friends

Ben Bucksch ben.bucksch at beonex.com
Wed Jan 16 19:00:25 UTC 2013


On 16.01.2013 16:00, Andrew Sutherland wrote:
> On 01/15/2013 10:55 PM, Joshua Cranmer wrote:
>> To my knowledge, Gaia runs in a content-privileged JS environment, 
>> which means it can't use nsISocketTransport (or any other XPCOM 
>> utilities).
>
> Yes, XPCOM is out of reach, but we create exciting new APIs to expose 
> the things we need.  This is mozTCPSocket:
> http://mxr.mozilla.org/mozilla-central/source/dom/network/src/TCPSocket.js 
>
>
> mozTCPSocket can also do the electrolysis thing, however, so 
> unfortunately things are somewhat more complicated than just changing 
> that JS.  Specifiically:
> http://mxr.mozilla.org/mozilla-central/source/dom/network/src/TCPSocketChild.cpp 
>
> http://mxr.mozilla.org/mozilla-central/source/dom/network/src/TCPSocketParent.cpp 
>
> http://mxr.mozilla.org/mozilla-central/source/dom/network/src/TCPSocketParentIntermediary.js 
>
>
> Originally, I think my Jetpack prototypes has startSSL support in 
> their fork, but the person working the bug did not propagate that over 
> and asked to have that be a follow-up so we could land sooner. They 
> never got around to the follow-up and then it wasn't getting marked as 
> a blocker so it couldn't be worked by anyone. Once things went 
> out-of-process and we got the extra gunk, the effort required also 
> went up a bit more.

(Something is going horribly wrong with the world when network APIs are 
underneath DOM, but whatever...)

FWIW, the startSSL method in XPCOM seems to be sync. When it should be 
async. So, whatever you want should work out. It's just a single 
function call from you to NSS, so that shouldn't be difficult to get 
into any API.

Maybe that's just me, but I think STARTTLS should be on the hard 
requirements list of any mainstream email client today. (It's good that 
you at least won't use unprotected connections.) If you don't support 
STARTTLS, you won't get configs for many ISPs. It's not just AOL.

Ben



More information about the tb-planning mailing list