[jnigen] Config API update by liamappelbe · Pull Request #3516 · dart-lang/native

liamappelbe · 2026-08-04T03:41:11Z

Update JNIgen's config API to align it with the proposed FFIgen API.

Group classes, sourcePath, classPath, summarizerOptions, mavenDownloads, and androidSdkConfig into Input
Group import and hide into SymbolImports, which lives in Input.imports
OutputConfig renamed to Output, and still contains dartConfig -> dart and symbolsConfig -> symbols. Now also contains preamble, generateStubs, and format.
nonNullAnnotations and nullableAnnotations now grouped under NullabilityAnnotations.
Flip isExcluded to isIncluded.
Make a bunch of nullable fields non-null, if null is functionally identical to a default value.
For the fields where null is functionally different, document that. Also, document any other undocumented fields.

Sorry for the crazy number of changed files. It's mostly due to the version bump. There aren't any real bindings changes. Important files are:

pkgs/jnigen/lib/src/config/config_types.dart, containing the new API
Any of the generation scripts, eg pkgs/jni/tool/generate_jni_bindings.dart, if you want to see the new API in action

#2062

github-actions · 2026-08-04T03:41:57Z

Package publishing

If you have publishing permissions, you can use the links below to publish the changes after merging this PR.

Package	Version	Status	Publish tag (post-merge)
package:code_assets	1.3.0-wip	WIP (no publish necessary)
package:data_assets	0.20.0	already published at pub.dev
package:ffi	2.2.0	already published at pub.dev
package:hooks	2.2.0-wip	WIP (no publish necessary)
package:hooks_runner	1.6.2-wip	WIP (no publish necessary)
package:jni_flutter	1.0.2	already published at pub.dev
package:jni_util	1.0.0	already published at pub.dev
package:native_toolchain_c	0.19.4-wip	WIP (no publish necessary)
package:record_use	1.1.0	ready to publish	`record_use-v1.1.0`
package:swift2objc	0.3.0-wip	WIP (no publish necessary)
package:swiftgen	0.1.3	already published at pub.dev

Documentation at https://github.com/dart-lang/ecosystem/wiki/Publishing-automation.

github-actions · 2026-08-04T03:43:26Z

PR Health

Breaking changes ✔️

Package	Change	Current Version	New Version	Needed Version	Looking good?
jni_flutter	None	1.0.2	1.0.2	1.0.2	✔️
jni	None	1.0.3	1.0.3	1.0.3	✔️

This check can be disabled by tagging the PR with skip-breaking-check.

API leaks ✔️

The following packages contain symbols visible in the public API, but not exported by the library. Export these symbols or remove them from your publicly visible API.

Package	Leaked API symbol	Leaking sources

This check can be disabled by tagging the PR with skip-leaking-check.

dcharkes · 2026-08-06T09:50:26Z

-          classes: ['com.github.dart_lang.jnigen.simple_package'],
-          sourcePath: [Uri.directory(sourcePath)],
+          input: Input(
+            classes: ['com.github.dart_lang.jnigen.simple_package'],


Helicopter view level question: We can't put this as include on visitClass right? Because this goes into the summarizer which is a different stage in the pipeline.

I think the readme of both FFIgen and JNIgen would benefit from mermaid diagrams showing the pipeline and what in the config is configuring which part of the pipeline. (Feel free to file an issue to do this later. Mermaid diagrams are rendered on GitHub, I don't know about pub, and probably not in dartdoc either.)

Another super high level question, should options only passed to the summarizer be nested in the summarizer options? That might make it more clear for users.

I think the readme of both FFIgen and JNIgen would benefit from mermaid diagrams showing the pipeline and what in the config is configuring which part of the pipeline. (Feel free to file an issue to do this later. Mermaid diagrams are rendered on GitHub, I don't know about pub, and probably not in dartdoc either.)

Done #3527

Another super high level question, should options only passed to the summarizer be nested in the summarizer options? That might make it more clear for users.

Hmmm. Tricky question.

Most of Input's remaining fields are only passed to the summarizer. Even MavenDownloads and AndroidSdk populate sourcePath and classPath, which are passed to the summarizer. The only one that is completely separate is imports.

I don't necessarily want to advertise to users that the summarizer is a separate tool we run. There's a backlog bug somewhere to port the summarizer to Dart. In that case, extraArgs, workingDirectory, and maybe even backend could be deleted.

classes, sourcePath, and classPath are way more common than any of the options in SummarizerOptions.

Just realized that in the FFIgen config, importType is a top level member of the config object.

With all that in mind, maybe a better design is to move imports to be directly under Config, and then delete SummarizerOptions and flatten all its options directly into Input?

I don't necessarily want to advertise to users that the summarizer is a separate tool we run.

Hm, the fact that we have a summarizer puts a specific burden on the user right now, listing the class names.

There's a backlog bug somewhere to port the summarizer to Dart.

I would imagine we completely drop the list of classes then.

Just realized that in the FFIgen config, importType is a top level member of the config object.

The imports in FFIgen are for types that are generated by other FFIgen runs right? That "feels" more like a different source of information than the header files. I think conceptually the input relates to the parsing stage of the native code.

So I'd probably keep all the summarizer options in input then, since that's the parsing stage. Including the list fo classes.

And maybe importType should be nested in another config that indicates it is about importing already generated bindings generated by the same tool.

So I'd probably keep all the summarizer options in input then, since that's the parsing stage. Including the list fo classes.

Do you mean keep it as it is in 5c24777 with the 3 old fields in SummarizerOptions, as it is in 11ccc34 with everything in Input. Or do you mean move classes, sourcePath, and classPath into SummarizerOptions?

I think 11ccc34 looks slightly more natural to me. We don't explicitly call out the summarizer in the name but basically all of those fields relate to the parsing stage and are used with the summarizer.

But if you feel strongly about it, I don't have a strong opinion.

I think that's the best approach too.

dcharkes · 2026-08-06T09:52:13Z

+          ],
+        ),
+        output: Output(
+          dart: DartCodeOutputConfig(


We can't use dot shorthands to skip typing the constructor name can we?

I think you might be able to do dart: .new(? Personally I don't find that more readable.

Ah right, yeah I also find that not readable 😅

dcharkes · 2026-08-06T09:53:06Z

+          ],
+        ),
+        output: Output(
+          dart: DartCodeOutputConfig(


We dropped Config suffix everywhere, this still has a config suffix.

I noticed some others still have the config suffix as well, e.g. AndroidSdkConfig.

goderbauer

LGTM

liamappelbe added 3 commits August 4, 2026 13:32

config API migration

9d2a73b

test and example migration

064b6a4

Migrate other monorepo generators

ca5ec79

github-actions Bot added package:jnigen package:jni package:jni_flutter labels Aug 4, 2026

liamappelbe added 11 commits August 4, 2026 13:58

version bump

54c326b

regen bindings

62d2de9

nit

f5446cb

Merge branch 'main' into jnigen_api_update

d5ee3ea

nit

60e2f6a

reorder fields to reduce diffs

f41e971

Make some config fields non-null

a66f252

fixes

80cbba1

Improve test output

a277ded

Fix test

9942136

Switch binding tests to Java 21

b6afc36

github-actions Bot added the type-infra A repository infrastructure change or enhancement label Aug 5, 2026

liamappelbe added 4 commits August 6, 2026 11:16

fix workflow

e3414f3

Flip isExcluded to isIncluded

2521a80

changelog

aaa64d9

Merge branch 'main' into jnigen_api_update

7c6406d

liamappelbe changed the title ~~WIP [jnigen] Config API update~~ [jnigen] Config API update Aug 6, 2026

liamappelbe marked this pull request as ready for review August 6, 2026 04:58

liamappelbe requested review from dcharkes and goderbauer and removed request for goderbauer August 6, 2026 04:58

zizmor

5c24777

liamappelbe added the skip-changelog-check label Aug 6, 2026

dcharkes reviewed Aug 6, 2026

View reviewed changes

liamappelbe added 4 commits August 7, 2026 10:22

Remove Config suffix

cfba39a

Reorganize inputs

20dc452

Merge branch 'main' into jnigen_api_update

ffdad7a

fmt

11ccc34

liamappelbe requested a review from dcharkes August 7, 2026 01:43

dcharkes approved these changes Aug 7, 2026

View reviewed changes

goderbauer approved these changes Aug 7, 2026

View reviewed changes

Uh oh!

Conversation

liamappelbe commented Aug 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Aug 4, 2026

Package publishing

Uh oh!

github-actions Bot commented Aug 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Health

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dcharkes Aug 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

goderbauer left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

liamappelbe commented Aug 4, 2026 •

edited

Loading

github-actions Bot commented Aug 4, 2026 •

edited

Loading

dcharkes Aug 7, 2026 •

edited

Loading