26 May 2012 01:45
Re: branch coverage for ternary/guard/default
Ned Batchelder <ned <at> nedbatchelder.com>
2012-05-25 23:45:46 GMT
2012-05-25 23:45:46 GMT
On 5/25/2012 6:31 PM, Jason Keene wrote:
I created a proof-of-concept of using settrace() to get bytecode-tracing rather than line tracing, but again, it's never been turned into a real tool, and in particular, it may not be possible to get useful reporting out of it. If you are interested, it's here: http://nedbatchelder.com/blog/200804/wicked_hack_python_bytecode_tracing.html
--Ned.
Hello,Most of the existing coverage tools are based on sys.settrace(), which only reports lines being executed, so it provides no visibility into the inter-line execution like you want to get. There are others based on AST transformations, but none of them have reached production-quality, they are experiments.
Is there a tool that can do branch coverage testing of ternary/guard/default expressions. Eg:'happy path' if mostly_true() else 1/0
mostly_true() or 1/0
mostly_false() and 1/0
So if the test suite doesn't hit the 1/0's then it will indicate partial coverage. I'm new to branch coverage testing and testing python in general so sorry if this is a newb question.
Jason
I created a proof-of-concept of using settrace() to get bytecode-tracing rather than line tracing, but again, it's never been turned into a real tool, and in particular, it may not be possible to get useful reporting out of it. If you are interested, it's here: http://nedbatchelder.com/blog/200804/wicked_hack_python_bytecode_tracing.html
--Ned.
_______________________________________________ testing-in-python mailing list testing-in-python <at> lists.idyll.org http://lists.idyll.org/listinfo/testing-in-python
RSS Feed