Closed jensmaurer closed 2 years ago
LEWG in Kona: Forward P1393 to LWG for C++20. SF F N A SA 4 7 3 7 7
Forward P1393 to LWG for a to-be-formed Executors TS. SF F N A SA 3 4 13 3 7
Forward P1393 to LWG for C++Next. SF F N A SA 11 7 4 6 2
Should be seen by LWG, no sooner than Belfast. Other Executors-related pieces will go to LEWG starting in Cologne for the C++23/C++Next cycle.
@tituswinters has asked that EWG see the paper.
EWG Prague Thursday afternoon: we saw the paper and will take a bit more time to understand what this paper does and why it does it in that particular way. We'll come back to it.
EWG telecon on May 21st 2020:
Properties are part of executors and are therefore one of the top items in the Bold Plan in P0592. Our main goal was to look at what's nominally a Library proposal and understand what, if anything, Language should do. In other words: don't design in isolation (neither Library nor Language).
One action item came out of this: Daisy to start a discussion with Daveed, Eric, Andrew, and Ville, regarding reflection + properties.
We didn't identify major parts of the proposal which should be done through Language facilities. We did, however, discuss a few places where better Language support might simplify properties / executors, and change what the Library parts look like:
None are blocking properties at the moment.
We took one straw poll, not as a binding decision but rather to understand what those present for the discussion thought.
Do we understand properties and think that specifying them purely in library is the right approach?
SF | F | N | A | SA |
---|---|---|---|---|
5 | 5 | 8 | 0 | 0 |
We'll revisit properties when EWG is next able to take binding decisions.
Initial LWG small group feedback provided to authors requesting a variety of updates.
Barry's P2279 will be in the coming mailing, "We need a language mechanism for static polymorphism", and might be related to this.
Poll: We understand properties and think that specifying them purely in library is the right approach.
SF | F | N | A | SA |
---|---|---|---|---|
1 | 4 | 14 | 4 | 2 |
Poll outcome: ❌ no consensus, participants do not understand properties, and might want to specify executors' customizations mechanism as a language feature.
Salient comments:
@JeffGarland given that Library Evolution is no longer pursuing P0443, but is instead moving forward with P2300, I don't think this should be in LWG's queue anymore.
P1393R0 A General Property Customization Mechanism (Daisy Hollman, Chris Kohlhoff, Bryce Lelbach, Jared Hoberock, Gordon Brown, Michał Dominiak)