Re: [myhdl-list] Best way to think in test driven development using MyHDL coming from VHDL?

[Christopher F. <chr...@gm...>]

On 7/15/2014 9:51 AM, André Prado wrote:
<snip>
>
> Is this the best way to do it?
>
> I am evaluation it by running:
>
> py.test tb_cordic.py -v
>
> It does work but I don't know if it's the best way to do it,

I don't know if there is a "best", many like to
use py.test, less scaffolding other prefer the
Python included unittest.

The great thing, MyHDL is Python - we can use
the Python references to investigate [1][2].

I haven't reviewed you code fully but I tend to
use a similar approach (that is use py.test),
I use this template

     def testbench(args):
         # all dut port instantiations
         clock = Signal(bool(0))
         reset = ResetSignal(0, active=0, async=True)
	# ...

         def _test():
             tbdut = # ...

             return tbdut, tbstim, ...

         Simulation(traceSignals(_test)).run()

I use the inner function because I often want to
view my testbench signals in the waveform and it
also allows multiple tests to be defined using
the same port definitions and arguments.

Typically, I use argparse to fill the test arguments,
this might be things like number of loop iterations
etc.

>
> I also have two functions doing almost the same thing, (random_bench and
> fft_bench), this violates the DRY mantra, but I could'nt find a way to do
> it generic enough.
>
> Taking all the DRY stuff (that also appears on my tests) that I messed up
> while testing, any good tips? Am I in the right path?

You could use the inner function to define your
"stimulus" and have the generic generators in the
outer function.  This will allow you to reuse
definitions and generic generators:

     def testbench():
         # signal instantiations

         @always(delay(3))
         def tbclk():
             clock.next = not clock

         @instance
         def tbmon():  #monitor
             # ...

         def _test_fft():
             tbdut =  # dut needs to be in the func to be traced
             # ...
             @instance
             def tbstim()
                 # ...
             return tbclk, tbmon, tbdut, tbstim, ....

        def _test_rand():
             tbdut = ...
             @instance
             def tbstim():
             # ...
             return tbclk, tbmon, tbdut, tbstim, ...


     if args.run == 'fft':
          tb = traceSignals(_test_fft)
     else:
          tb = traceSignals(_test_rand)
     Simulation(tb).run()


Regards,
Chris


[1] http://docs.python-guide.org/en/latest/writing/tests/
[2] http://www.ibm.com/developerworks/aix/library/au-python_test/