patch - tests in bitfcns

[Robert T. Short]

Thread switch was understood. 

Bob is still learning.   I guess I worry more about code coverage than 
the number of tests, but still, your thoughts make sense.  Since I 
didn't write the code and don't completely understand it I don't really 
imagine I got the code coverage thing all that well either.

I will update the bitfcns tests and recheck against MATLAB.  Probably 
take a couple of days though.  Are there any other 
permutation/combinations that should go in there?


Bob

John W. Eaton wrote:
> On 15-Jul-2009, John W. Eaton wrote:
>
> | On 15-Jul-2009, Robert T. Short wrote:
> | 
> | | Thanks.  My bugs are better than octave bugs.  Must have done something 
> | | dumb in the translation back from MATLAB to octave.  I am still working 
> | | on automating the whole process, so I suppose glitches are inevitable.
> | 
> | It probably depends on how closely related the tests are.  For
> | example, do all of the calls to assert from the first test block in
> | conv.m test the same thing, or are they really testing differennt
> | things?  If they are testing different things, then maybe they should
> | be separate tests?  I'm not sure what is best.
>
> Oops, this reply was really intended for the other thread in response
> to the question about whether separate tests are better.
>
> jwe
>
>
>