[asterisk-dev] About asterisk development plans

[Peter Beckman]

On Wed, 18 Mar 2009, Steve Murphy wrote:

> But asterisk is quite challenging. To do it right, I picture
> electro-mechanical devices that will pick up a handset and dial physical
> analog phones, supplying standard sound sequences, and recording sound
> into files.

  I think you're getting way too far ahead.  That sort of thing is 2-3 years
  out if an automated testing system were to start being built now.

  First start off by making sure each library function call is doing the
  "right thing" given a variety of inputs.

  Then once tests are written for the core libraries, you can get into some
  more advanced stuff.

  Write a test for the Asterisk Manager.  I'm having a frustration now where
  Action: Queues returns an EOL after every queue, so for that function, I
  have to find 2 EOL's in a row, while for other functions, a single EOL is
  enough, or for Commands --END COMMAND-- identifies the end of a command.
  The trouble is, if this changes in the next version, my current
  implementation breaks.  With a test of Asterisk Manager, changes would
  show up in such tests, giving you some insight as to what you might change
  (end ALL Actions with 2 EOLs or a simple Response: End) to make sure that
  people using Asterisk Manager won't be surprised by the next release that
  adds a command with an inconsistent end of response identifier.  It's a
  simple problem to fix, but extremely frustrating when it breaks.

  Now you can use the Asterisk Manager (now that you know the core libraries
  and AstMan are functioning properly) to originate calls, transfer calls,
  etc to test larger functionality.

  This leads to being able to write a test for every CLI command, to make
  sure it gives the correct result with a given scenario.  That might
  require having a few active SIP calls and a given dialplan, which you can
  now originate using Asterisk Manager.

  I use Asterisk for SIP only, and I'm guessing that a fair amount of people
  do.  Focusing the core of automated testing on easily automatable stuff
  will build you a foundation.  The foundation will mature as more people
  build tests and find frustrations, leading to a more rubust testing
  platform.  It is then that more complex testing, such as analog lines,
  Digium and 3rd party devices, etc will be an easier challenge to face.

---------------------------------------------------------------------------
Peter Beckman                                                  Internet Guy
beckman at angryox.com                                 http://www.angryox.com/
---------------------------------------------------------------------------
-------------- next part --------------
_______________________________________________
--Bandwidth and Colocation Provided by http://www.api-digital.com--

asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-dev