dubslow

> Hmm.. I have to say I still don't quite get the problem. Like, this is what I think of when I think "asyncio.gather" to try to demonstrate what I...

Let me try to summarize. Nurseries are not yet complete implementations of structured concurrency: the fact that all these `gather`-like recipes use closures (at least in their shortest form they...

> In regards to creating a function for each task, that works and is a quite powerful approach, but it does have the problem that these functions are inherently specialised...

Ah, very nice. That is indeed approximately the functionality that a nursery should implement by default. Ultimately I think it shouldn't be necessary to store some new variable from within...

> Yes a nursery could save the results of all those tasks. But how do you access them later? Sequentially? (By task start or by task stop?) By some key?...

> Which sequence – task start or task exit? Considering that the former is nonsensical (saving a value that doesn't exist yet), the latter. > What about nurseries that run...

Well "enforce" is always a strong word, Python users can easily shoot themselves in the foot with any code if they try hard enough, but in this case I just...

the simple LTC has proven to be considerably less good (but it was good to have tested it ofc), so i vote we stop that second LTC and approve this...

I've pushed a new version with cj's constant names, and expanding the comment based on peregrine's concerns

> cutoff priority that's from existing master, and wasn't my contribution. not related to this PR, tho perhaps you think this PR is a good place to fix this issue...