Create Static Enough Metaprogramming proposal #4374

mraleph · 2025-05-14T11:41:45Z

This moves content from #4271 into a markdown file in the repository to make discussion and revisions easier.

I have incorporated some of the feedback from discussions on the issue - but I continue to maintain focus on this as a toolchain feature. I have added some remarks that analyzer can't constant fold everything anyway because it does not have access to the compilation environment.

I would like to collect a few rounds of feedback and then rejuvenate the prototype implementation to get something experimental working across all platforms in the SDK so that we can have an idea of how well this could work in a real world.

lrhn · 2025-05-14T11:43:53Z

working/4271 - static enough metaprogramming/proposal.md

+introspect program structure which are _required_ to execute in compile time
+toolchain supports that. These two together should give enough expressive power
+to solve a wide range of problems where metaprogramming is currently wanted. cc
+@dart-lang/language-team


No cc needed here.

lrhn · 2025-05-14T11:43:54Z

working/4271 - static enough metaprogramming/proposal.md

@@ -0,0 +1,1222 @@
+tldr: I propose we follow the lead of **D**, **Zig** and **C++26** when it comes


I'd give this file a header.

A title line at the top.

An author line: Author: @mralpeh.

Consider a version line: version: 1.0 and a version section/changelog at the end, to track changes.

eernstg

LGTM. (Very interesting indeed!)

I could suggest the addition of 'a' or 'the' in many locations, but decided that this wouldn't be the top priority at this time.

eernstg · 2025-05-15T11:24:13Z

working/4271 - static enough metaprogramming/proposal.md

+to metaprogramming. We introduce an optional (toolchain) feature to force
+compile time execution of certain constructs. We add library functions to
+introspect program structure which are _required_ to execute in compile time
+toolchain supports that. These two together should give enough expressive power


Perhaps?:

Suggested change

toolchain supports that. These two together should give enough expressive power

if the toolchain supports that. These two together should give enough expressive power

Also, 'at compile time' seems to be more common than 'in compile time'.

eernstg · 2025-05-15T11:25:15Z

working/4271 - static enough metaprogramming/proposal.md

+# History of `dart:mirrors`
+
+In the first days of 2017 I have written a blog post ["The fear of
+`dart:mirrors`"][the-fear-of-dartmirrors] which contained started with the


Suggested change

`dart:mirrors`"][the-fear-of-dartmirrors] which contained started with the

`dart:mirrors`"][the-fear-of-dartmirrors] which started with the

eernstg · 2025-05-15T11:25:53Z

working/4271 - static enough metaprogramming/proposal.md

+> fog of uncertainty and marked as _Status: Unstable_ in the documentation -
+> even though APIs have not changed for a very long time.
+
+In 2017 type system was still optional, AOT was a glorified


Suggested change

In 2017 type system was still optional, AOT was a glorified

In 2017 the type system was still optional, AOT was a glorified

eernstg · 2025-05-15T11:27:41Z

working/4271 - static enough metaprogramming/proposal.md

+_"ahead-off-time-JIT"_, and the team maintained at least 3 different Dart
+front-ends (VM, dart2js and analyzer). Things really started shifting with Dart
+2 release: it had replaced optional types with a static type system and
+introduced _common front-end (CFE)_ infrastructure to be shared by all backends.


Suggested change

introduced _common front-end (CFE)_ infrastructure to be shared by all backends.

introduced a _common front-end (CFE)_ infrastructure to be shared by all backends.

eernstg · 2025-05-15T11:29:09Z

working/4271 - static enough metaprogramming/proposal.md

+and necessity to use _AOT-compilation_ for deployment.
+
+Dart 1 was all in on dynamic typing. You don't know what a variable contains -
+but you can do anything with it. Can pass it anywhere. Can all any methods on


Perhaps?:

Suggested change

but you can do anything with it. Can pass it anywhere. Can all any methods on

but you can do anything with it. Can pass it anywhere. Can call any methods on

eernstg · 2025-05-15T11:34:10Z

working/4271 - static enough metaprogramming/proposal.md

+>
+> - [Reflection Analysis for Java][paper-java-suif-reflection]
+> - [Understanding and Analyzing Java Reflection][paper-java-reflection-2019]
+> - [Challenges for Static Analysis of Java Reflection – Literature Review and Empirical Study](paper-java-reflection-challenges)


Just guessing:

Suggested change

> - [Challenges for Static Analysis of Java Reflection – Literature Review and Empirical Study](paper-java-reflection-challenges)

> - [Challenges for Static Analysis of Java Reflection – Literature Review and Empirical Study][paper-java-reflection-challenges]

eernstg · 2025-05-15T11:37:52Z

working/4271 - static enough metaprogramming/proposal.md

+> [!NOTE]
+>
+> Another exploration similar in nature was
+> (go/const-tree-shakeable-reflection-objects)[http://go/const-tree-shakeable-reflection-objects].


Perhaps?:

Suggested change

> (go/const-tree-shakeable-reflection-objects)[http://go/const-tree-shakeable-reflection-objects].

> [go/const-tree-shakeable-reflection-objects](http://go/const-tree-shakeable-reflection-objects).

eernstg · 2025-05-15T11:41:23Z

working/4271 - static enough metaprogramming/proposal.md

+very small and excludes a lot of expressions which feel like they should
+actually be included. It just feels wrong that `const x = [].length` is invalid
+while `const x = "".length` is valid. For some seemingly arbitrary reason
+`String.length` is the only blessed property which can't be accessed in a


Suggested change

`String.length` is the only blessed property which can't be accessed in a

`String.length` is the only blessed property which can be accessed in a

jakemac53 · 2025-05-15T17:26:37Z

working/4271 - static enough metaprogramming/proposal.md

+> Another exploration similar in nature was
+> (go/const-tree-shakeable-reflection-objects)[http://go/const-tree-shakeable-reflection-objects].
+
+Fundamentally both of these approaches were dead ends and Web applications


Fwiw, I wouldn't say that the const reflection proposal was a dead end at all, we just simply didn't end up doing it but there wasn't any sort of blocker or anything that I remember encountering.

jakemac53 · 2025-05-15T17:38:31Z

working/4271 - static enough metaprogramming/proposal.md

+> This proposal does not suffer the same issue, even if we decide to support
+> `@konst` evaluation in these tools: as `@konst` actually provides a syntactic
+> marker which partitions the program. Only bodies of methods with `@konst`
+> parameters need to be included into the outline.


Can these functions only invoke other @konst functions? This was a big problem with enhanced const - it quickly ends up being that all functions end up being marked as "const" just because some other const function completely unrelated to your function wants to invoke them, and your function could be const so you say sure, but now all these functions end up in outlines.

I think you need something like @konst imports so it is defined at the usage site and not the definition site which functions can be called by @konst functions.

jakemac53 · 2025-05-15T17:48:01Z

working/4271 - static enough metaprogramming/proposal.md

+Dart source code generation or which are supported by macro systems in other
+programming languages. For example the following capabilities are out of scope:
+
+- injecting new declarations into the program;


This definitely does drastically simplify the proposal - but it also makes it a lot less useful. Ultimately trying to accomplish this was the downfall of the original macros proposal though.

But, I think it is important to look at the original motivations for macros, and which of those use cases are covered by this.

Ultimately, I think you can end up with some fairly decent solutions for automatic encoding/decoding, equality, toString, etc. But not copyWith or constructors due to the signatures being dependent on the shape of the class (you could generate the bodies of these but that's only half the boilerplate).

Edit: I see you have an interesting copyWith idea using records, that is fairly reasonable actually.

jakemac53 · 2025-05-15T17:50:00Z

working/4271 - static enough metaprogramming/proposal.md

+> }
+> ```
+>
+> Though usability of this method will be questionable as there will be good


Suggested change

> Though usability of this method will be questionable as there will be good

> Though usability of this method will be questionable as there will be no good

jakemac53 · 2025-05-15T17:52:45Z

working/4271 - static enough metaprogramming/proposal.md

+- Average code size overhead per class: 270 bytes
+- Average JIT kernel generation overhead per-class: 0.2ms (cost of producing
+  specialized functions using Kernel-to-Kernel AST transformation)
+- Average AOT compilation overhead per-class: 2.6ms


What is the baseline?

jakemac53 · 2025-05-15T17:53:58Z

working/4271 - static enough metaprogramming/proposal.md

+code:
+
+```dart
+mixin DataClass<@konst T> {


This would be 🔥 (and help a lot with the reduction of boilerplate)

kevmoo · 2025-05-15T20:08:52Z

working/4271 - static enough metaprogramming/proposal.md

+
+This requires the definition of constant expression to be expanded to cover a
+significantly larger subset of Dart than it currently includes. Such feature
+does however exist in other programming languages, most notably **C++**, **D**,


Suggested change

does however exist in other programming languages, most notably **C++**, **D**,

do however exist in other programming languages, most notably **C++**, **D**,

kevmoo · 2025-05-15T20:10:58Z

working/4271 - static enough metaprogramming/proposal.md

+}
+```
+
+Note that all methods are annotated with `@konst` so if compiler supports


Suggested change

Note that all methods are annotated with `@konst` so if compiler supports

Note that all methods are annotated with `@konst` so if the compiler supports

mraleph added 2 commits May 14, 2025 13:33

Initial version of the proposal

72d111b

Tweak wording a bit

43f4ec2

mraleph requested review from munificent, jakemac53, lrhn, leafpetersen and eernstg May 14, 2025 11:41

lrhn approved these changes May 14, 2025

View reviewed changes

eernstg approved these changes May 15, 2025

View reviewed changes

jakemac53 reviewed May 15, 2025

View reviewed changes

kevmoo reviewed May 15, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Create Static Enough Metaprogramming proposal #4374

Create Static Enough Metaprogramming proposal #4374

mraleph commented May 14, 2025

lrhn May 14, 2025

lrhn May 14, 2025

eernstg left a comment

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

eernstg May 15, 2025

jakemac53 May 15, 2025

jakemac53 May 15, 2025 •

edited

Loading

jakemac53 May 15, 2025 •

edited

Loading

jakemac53 May 15, 2025

jakemac53 May 15, 2025

jakemac53 May 15, 2025 •

edited

Loading

kevmoo May 15, 2025

kevmoo May 15, 2025

		@@ -0,0 +1,1222 @@
		tldr: I propose we follow the lead of D, Zig and C++26 when it comes

	toolchain supports that. These two together should give enough expressive power
	if the toolchain supports that. These two together should give enough expressive power

	`dart:mirrors`"][the-fear-of-dartmirrors] which contained started with the
	`dart:mirrors`"][the-fear-of-dartmirrors] which started with the

	In 2017 type system was still optional, AOT was a glorified
	In 2017 the type system was still optional, AOT was a glorified

	introduced _common front-end (CFE)_ infrastructure to be shared by all backends.
	introduced a _common front-end (CFE)_ infrastructure to be shared by all backends.

	but you can do anything with it. Can pass it anywhere. Can all any methods on
	but you can do anything with it. Can pass it anywhere. Can call any methods on

	> - [Challenges for Static Analysis of Java Reflection – Literature Review and Empirical Study](paper-java-reflection-challenges)
	> - [Challenges for Static Analysis of Java Reflection – Literature Review and Empirical Study][paper-java-reflection-challenges]

	> (go/const-tree-shakeable-reflection-objects)[http://go/const-tree-shakeable-reflection-objects].
	> [go/const-tree-shakeable-reflection-objects](http://go/const-tree-shakeable-reflection-objects).

	`String.length` is the only blessed property which can't be accessed in a
	`String.length` is the only blessed property which can be accessed in a

	> Though usability of this method will be questionable as there will be good
	> Though usability of this method will be questionable as there will be no good

	does however exist in other programming languages, most notably C++, D,
	do however exist in other programming languages, most notably C++, D,

	Note that all methods are annotated with `@konst` so if compiler supports
	Note that all methods are annotated with `@konst` so if the compiler supports

Create Static Enough Metaprogramming proposal #4374

Are you sure you want to change the base?

Create Static Enough Metaprogramming proposal #4374

Conversation

mraleph commented May 14, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

eernstg left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jakemac53 May 15, 2025 • edited Loading

Choose a reason for hiding this comment

jakemac53 May 15, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jakemac53 May 15, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jakemac53 May 15, 2025 •

edited

Loading

jakemac53 May 15, 2025 •

edited

Loading

jakemac53 May 15, 2025 •

edited

Loading