[Twisted-Python] Trial regression

Andrew Bennetts andrew-twisted at puzzling.org
Sun Sep 4 19:36:22 MDT 2005


On Mon, Sep 05, 2005 at 11:19:51AM +1000, Jonathan Lange wrote:
> Actually, it'd make things go quicker if you could help me write a
> unit test for this behaviour.
> 
> Also, I'd appreciate advice on ways to avoid further regressions yet
> still maintain a good pace of development.

The low-tech answer would be to diff the output of "bin/trial -R twisted"
using trial from trunk vs. trial from your branch (for both stdout and
stderr, of course).  That will at least catch any differences in the major
project people care about.

I expect that as you increase test coverage, regressions will be less and
less common.

-Andrew.





More information about the Twisted-Python mailing list