Morgan :)

Consider running macros in rounds instead of phases

9

Investigation for #3706 suggests an alternative way to structure how macros run that we should consider. The problem is how macro output affects other macros: what can macros see, and...

Quantify the cost of merging augmentations into the analyzer and CFE data models

2

Thanks to @scheglov for [documenting](https://github.com/dart-lang/language/blob/main/working/macros/host-notes.md) how the analyzer hosts macros, this makes the discussion much more concrete :) For performance, the critical piece we can dig into from here seems...

Consider generalizing macros over applications and annotations

12

A difference that cropped up naturally in the `dart_model` exploration, and seems worth considering independently of that work. _Generalize over applications:_ Consider having one macro "runtime instance" handle multiple applications....

Macro annotation information in yaml files next steps

3

Per #3728 there is a plan to pull macro data--information about which annotations trigger macro execution, and exactly what they trigger--into yaml files and from there into `package_config.json` so it's...

"Merge to source" for macros that only need Phase 3

26

Here's an alternative to Phase 1+2. @jakemac53 @johnniwinther @scheglov It's not new: it's the recommended way of doing codegen in Java at Google, and I tried to do it for...

Strategy for launching augmentations

6

A small pile of questions to resolve here, that are important in considering macros timelines. 1) There is a possibility that we launch a single macro, e.g. the JSON macro,...

Next steps for `dart_model` prototype integration

16

We have a bunch of stuff to break--see milestone--and it's hard enough already without also worrying about breaking rolls / analyzer users. WDYT? @jakemac53 @scheglov

Pick a supported serialization format for host<->macro communication

26

The [dart_model exploration](https://github.com/dart-lang/language/tree/main/working/macros/dart_model) currently uses JSON but there is no particular reason to think that's a good choice :) Even if we don't follow the `dart_model` direction, it seems likely...

In case this isn't tracked elsewhere; we should follow up on turning https://dart-review.googlesource.com/c/sdk/+/232682 into an implementation that doesn't require "cast to dynamic", or removing, as appropriate.

Design for introspection on macro metadata / annotations

30

Related: #3522 @johnniwinther @jakemac53 is there an issue already open for the investigation Johnni's been doing into macro metadata? I couldn't find one, so here's a fresh one :) and...