[asterisk-users] Discussion: Test platform

Olle E. Johansson oej at edvina.net
Thu May 5 10:07:29 CDT 2011


> 
> Here is the thing, there is nothing stopping 'the community' today from doing this.  In fact, we already have a testsuite [1] in place, running each subversion commit and producing results for the last year.  But this is only one type of testing; automated, we also have unit tests built into Asterisk that run too (EG: a unit test to parse SIP URI). Again, each subversion commit we run the tests and validate results.

I think we should make it more clear and give examples on how we can extend the test platform to test functionality in our own platforms - our dialplans and channel drivers. If we did that, more people would use the test toolkit and work with it daily.

> 
> There is still lots of work that needs to be done though. More test plans and test cases to be added, more code to be written and libraries added, getting more people involved in testing Asterisk Release Candidates (RCs) or patches on the issue tracker.
> 
> That is the hardest part, getting people involved.  Sure it is easy to say Asterisk is not stable, not production ready or it crashes all the time; fair enough but we have tools in place to help resolve that. Just in this thread alone I don't believe one person has answered the call of Olle to volunteer time to help maintain Asterisk 1.4 (if I am incorrect please speak up, I must have missed your name). Additionally, this almost exact point was raise on the asterisk-dev mailing list in 2009 [1] (a great read BTW, lots of great ideas) however due to the lack of interest it did not go to far.
If you go even further back, Russell and I had a branch where we started some early work many, many years ago. We're asterisk-dinosaurs in that respect... I am very happy that we now, eons later, have a test toolkit. It's lightyears ahead of what we discussed or dreamed of back then. And it has helped a lot in catching stuff.

> 
> So how can we fix this?  How can we get more people involded?  What makes projects like FedoraTesting[3] and DebianTesting[4] popular?  How can the Asterisk project reproduce their success?
Give them something that tests their own setup as well as test the Asterisk in the core.

> 
> As I've said before, I'm more then willing to help with answering questions about the testsuite or reviewing code that people want to get merged in.  We also have an IRC channel, #asterisk-testing available for people to join, ask question, idle, lurk, etc, or if you want to reply to this thread, feel free.  But get involved! :)

Absolutely - we need people that test the new bugs that  developers invent :-)

/O


More information about the asterisk-users mailing list