Danielle Madeley

Lettuce users are encouraged to take a look at [Aloe](https://github.com/aloetesting/aloe), which grew out of an attempt to port Lettuce to Python 3 and fix up the bugs in the parser.

I endorse mock. Also if you're not keen on six, have a look at python-future.

I've got a branch that extensively refactors the core of Lettuce (#429), I am taking the time as part of this work to try and makes things as Python3/unicode clean...

You can get this from the Xunit/Subunit output.

Hi, this is wrong because it breaks the outputters. In the new-parser branch we have before/after.each_example which implements what you want without breaking the API. You could backport that here.

Hi, it's still wrong, because it still breaks the outputters, you need both before/after scenario and before/after example.

This is still the wrong approach. It breaks the API. Have a look at how this has been approached in the new-parser branch.

This is fixed in the infoxchange/new-parser branch which is planned to become lettuce 0.3.

I've wanted the ability to rerun the broken tests from the last run. I can help you implement something like that off the new-parser branch (which is much clearer).

I think you're going to need to add support in lettuce core. The hook architecture wouldn't let you do what's needed here I don't think. Have a look in Scenario.run()