Author yselivanov
Recipients Petter S, asvetlov, pdxjohnny, r.david.murray, yselivanov, zach.ware
Date 2018-03-06.19:26:14
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1520364374.87.0.467229070634.issue32972@psf.upfronthosting.co.za>
In-reply-to
Content
> How is a separate base class better? :)

It's very explicit that way.

Also, I personally subclassed TestCase in many of my projects specifically to add async support.  To do that you have to use a metaclass to scan class' namespace for 'async def' functions.

Currently, unittest.TestCase doesn't have a metaclass.  If you add one to it, it might break all packages that were subclassing TestCase with a metaclass.


> If you accidentally add `async` to the front of a test method in a TestCase-derived test class, you get mostly-silent success with an easily-ignored warning about a coroutine not being awaited.

Well, what if you use trio or curio?  You add an 'async' keyword and get a cryptic error that some framework internals just broke.

So I'm strong -1 on the coroutine_runner attribute idea.
History
Date User Action Args
2018-03-06 19:26:14yselivanovsetrecipients: + yselivanov, r.david.murray, asvetlov, zach.ware, pdxjohnny, Petter S
2018-03-06 19:26:14yselivanovsetmessageid: <1520364374.87.0.467229070634.issue32972@psf.upfronthosting.co.za>
2018-03-06 19:26:14yselivanovlinkissue32972 messages
2018-03-06 19:26:14yselivanovcreate