[eispice] Re: behavior test case problems
- From: Charles Eidsness <charles@xxxxxxxxxxxxxxxxxxxxx>
- To: eispice@xxxxxxxxxxxxx
- Date: Thu, 28 Dec 2006 12:49:53 -0500
Hi Jon,
I've tried to improve the PyB related error messages for the next
release. I think that was the root of the issues getting that model to
work. It's tough to debug when every minor syntax error returns the
same, incorrect error text.
The next release should be ready in a week or two. I've been working on
a simple eispice IDE using PyQt (which will be packaged separately) in
hopes of attracting more Windows users. I've been delaying the next
release until the IDE is ready, but now I might release eispice earlier.
I've made some performance enhancements to the PyB device today (it's
about 300% faster now) that will probably be useful for you if you're
going to be writing more complicated PyB models.
There's a set of tests in the test/pass directory of the source
distribution that I've been using for regression testing, there's a
test.sh script in the root directory of the source distribution that
runs them all. Most of the tests are pretty general and simplistic, that
area could definitely use some work. As new things are discovered, like
the LC-Ladder Bug that Thomas uncovered or your PyB model I've been
adding tests for them. That brings up a good point; if anyone has any
circuits that they would like me to add to the test suite, to ensure
they will work with every new eispice release I'm happy to add and run them.
Cheers!
Charles
Jon Choy wrote:
> Charles,
>
> I'm out of town at the moment; but I do see the email trail with the
> EISPICE deck I sent and I think I added that resistor in parallel because I
> was trying to debug why I initially couldn't get a DC solution - I was
> worried about a divide by zero in the circuit matrix. Sorry for the
> confusion, I shouldn't have sent a bunch of test cases encapsulated into one
> netlist. With your next release, I will try this exercise again. I think
> last week I tried some of your suggestions by adding spaces for your parser
> and wasn't able to get a DC solution. If I still have problems later when I
> get back into town, I may want to see your deck so I can DIFF my deck with
> yours.
> On another note, I don't know if you are putting together a list of test
> cases or have a list of test cases as part of your QA process. In the
> industry that I work in the set of tests are called regressions or a
> regression suite. As you make changes and additions to EISPICE, you will
> want to make sure that as you fix your tool for one thing that it doesn't
> break other tests that should have not been affected. Just a thought...
> Take it for what it's worth.
>
------------------------------------------------------------------
To unsubscribe from the eispice list send an email to:
eispice-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
Other related posts:
