[Opensim-dev] LSL test suite ideas

[Mike Mazur]

Hi,

On Mon, 30 Jun 2008 15:41:10 -0500
krtaylor <krtaylor at linux.vnet.ibm.com> wrote:

> I have been putting together a few ideas on how to include more tests 
> for LSL. I am thinking of putting together a set of scripts that
> would systematically test all the script functionality that is
> currently implemented.

Zaki and myself have just talked about this only yesterday. We thought
of writing a bot (pCampBot?) that could log in, create a prim with one
script. The script in its touch_start() event would call the one
function you want to test and spit out the result in chat. Presumably
the bot could verify the message against a predefined expected message,
and report on its success. (I just realized the state_entry() event
could be used as a setup routine.)

The downside to this is you're depending on much of the underlying
script engine implementation to be correct for the llFunction() under
test to return the correct result. So a test failure could indicate
either a bug in the function implementation or somewhere else in the
scripting engine.

Using a bot has a few benefits, such as being able to log in to
multiple grids with the same client and the ability to set up a
comprehensive automation network with email reports, etc.

Thoughts on this?

Mike