[erlang-questions] Re: testing asynchronous code

[John Hughes]

We've actually just written a paper about asynchronous testing, based on 
testing ejabberd (where the message delivery events are asynchronous). The 
paper talks about how to analyze a recorded trace, without falling into the 
twin traps of complex code or expensive analysis. I've put a copy online at

www.cs.chalmers.se/~rjmh/ast2010.pdf

if you're interested. (The paper will be published at AST 2010 in Cape Town 
in May).

John


----- Original Message ----- 
From: "Tim Fletcher" <mail@REDACTED>
To: <erlang-questions@REDACTED>
Sent: Monday, April 19, 2010 1:56 PM
Subject: [erlang-questions] Re: testing asynchronous code


>> Anyone have general strategies or advice for testing code that makes
>> heavy use of gen_server:cast? The best I could come up with was to
>> write _sync versions of all my functions, with gen_server:cast
>> replaced by gen_server:call, and then test those, but that seems both
>> clunky and fragile.
>
> How about using a "test spy" process to collect a trace of the cast
> messages and then make assertions against the resulting trace?
>
> Tim
>
> ________________________________________________________________
> erlang-questions (at) erlang.org mailing list.
> See http://www.erlang.org/faq.html
> To unsubscribe; mailto:erlang-questions-unsubscribe@REDACTED
>
>
>