[TIP] Sybil 1.0.0 Released!
chris at withers.org
Tue May 30 10:41:16 PDT 2017
On 26/05/2017 09:02, Harry Percival wrote:
> I did quite a lot of testing for my book about testing. If you're
> looking for inspiration for future features, the code is at:
Sounds like you could have used Manuel? And of course: Sybil, but my
time machine is unfortunately faulty...
> I wanted to be able to:
> - parse listings and incomplete listings
Not sure what this means?
> - handle state including multiple files,
Sybil supports state across examples within a file, cross-file state is
a little weird, no?
> including non-python code
Writing a parser/evaluator for this should be fairly trivial...
> - check that arbitrary terminal commands, if executed in the cumulative
> context of the listings so far, actually produce the output the book
> says they do (tests and their tracebacks mostly)
Heh, the example I use in Sybil's docs is a very basic version of this:
> I ended up with something that may be a little too specific to my use
> case, and a little overengineered, and spaghetti code full of the
> vestiges of abandoned dead end approached. A brief attempt to roll my
> own textwrap.textwrap for example.
I'm quite happy with Sybil's code, it was artisanal* after all.
> But still, end results are there. And of course yes there are tests for
> the tests of the tests in the testing book.
> Happy to talk anyone through any of the code if it's of interest.
If you want to refactor it to use Sybil, I'd be happy to help if you hit
> Otherwise please just pat me on the back and tell me how smart I am? ;-)
Could I just buy you a pint instead?
More information about the testing-in-python