Reintroduce default value for ignore_args argument of FixtureManager.getfixtureclosure

[nicoddemus]

Fix #11868

[YannickJadoul]

If I read this correctly, the order of arguments also changed in #11416.

Not my decision as to whether you want to make this backwards-compatible again with previous versions, but if you want to, the current PR likely isn't enough?

[bluetech]

@YannickJadoul you're right. I think the function changed too much (also in its semantics) that it's not really possible to maintain backwards compat for it. The situation with pytest-lazyfixtures is unfortunate since it seems it's abandoned without any way to transfer ownership. It probably needs a fork...

[nicoddemus]

@YannickJadoul you're right. I think the function changed too much (also in its semantics) that it's not really possible to maintain backwards compat for it.

In that case, it is probably better to not merge this at all, as will only add to the confusion?

[bluetech]

Yes, I think we shouldn't merge it.

[YannickJadoul]

The situation with pytest-lazyfixtures is unfortunate since it seems it's abandoned without any way to transfer ownership. It probably needs a fork...

Agreed, yes! That's why I explicitly stepped back from saying I'd like to see this change.
pytest-lazyfixtures will need more than just this PR, so I wouldn't consider it as a good reason to merge it.

[nicoddemus]

Thanks folks, closing this then. 👍

[RonnyPfannschmidt]

should we consider integrating lazy-fixtures in core - its in line with the fixture configuration details i have in mind

i already tried to reach out to the maintainer a while back and got no reply

[YannickJadoul]

should we consider integrating lazy-fixtures in core - its in line with the fixture configuration details i have in mind

i already tried to reach out to the maintainer a while back and got no reply

I don't have massive amounts of time, but if this isn't super urgent and can take e.g. a few weeks, I would be interested in helping out here!
And maybe @youtux as well? (see TvoroG/pytest-lazy-fixture#63)

In the end, lazy-fixures is not a massive amount of code. I'd just need to look into all the previous issues and discussions had before, and get some input on whether the current implementation is still the best way forward or whether there's a more stable way of implementing it.

[RonnyPfannschmidt]

if pytest introduces a native object to statically request the request for another fixture as parameter (with or without a reconfigrue of the fixture parameters) implementation may look different

[nicoddemus]

@RonnyPfannschmidt @YannickJadoul there's a detailed topic already in #11532, I suggest to move the discussion over there.