[Spread-users] where can I find documentation of the protocol for connecting to spread daemons?
alex at alexjacobson.com
Wed Nov 7 14:37:26 EST 2007
The java implementation blocks on callbacks, so a poorly behaved
MessageListener can block the whole app. Is that intended behavior
because you want catastrophic failure or is that just an oversight?
The alternative is to use some intermediary mailbox object that doesn't
hold up the callback engine.
Also, SpreadAuthInterface is confusing. Is authentication handled on a
separate connection or the same connection?
Jonathan Stanton wrote:
> We don't have a formal spec for the client-server protocol in Spread.
> Most of the client libraries wrap the C library libspread, so they don't
> need to know the actual protocol. There are a few other native
> implementations including the Java library that directly implement the
> network protocol. So the best documentation for the network protocol is to
> look at the C or java library code (some parts of the protocol are
> explained in the comments). The messages exchanged by the client and
> server are pretty straightforward and use a standard message format.
> Let me know if there is any specific information you want to know about
> On Tue, Nov 06, 2007 at 01:17:47PM -0800, Alex Jacobson wrote:
>> I'd like to write my own client in Haskell, but all I can find are other
>> examples. Is there a formal spec somewhere?
>> Spread-users mailing list
>> Spread-users at lists.spread.org
More information about the Spread-users