Lint check for private fields

[remcomokveld]

When using kotlin-codegen the annotation processor will crash if a JsonClass annotated class, or any of super classes, contains a private non-transient field.

With #728 it became possible to only enable the processor for certain compilation targets. The disadvantage of this is that a debug build might succeed while a release build fails.

To get earlier feedback on this during development it would be nice to have a lint check with quick-fixes to either mark a field as transient, or make it more accessible.

I created a gist here with the lint check and a test
https://gist.github.com/remcomokveld/a96a245cc21f3976b5c56b27b7384767

This would need an additional AAR artifact which contains just the lint jar, so not sure if it is feasible

[JakeWharton]

This isn't an Android library and therefore can't ship lint checks.

…

On Tue, Jan 1, 2019, 4:13 PM Remco Mokveld ***@***.*** wrote: When using kotlin-codegen the annotation processor will crash if a JsonClass annotated class, or any of super classes, contains a private non-transient field. With #728 <#728> it became possible to only enable the processor for certain compilation targets. The disadvantage of this is that a debug build might succeed while a release build fails. To get earlier feedback on this during development it would be nice to have a lint check with quick-fixes to either mark a field as transient, or make it more accessible. I created a gist here with the lint check and a test https://gist.github.com/remcomokveld/a96a245cc21f3976b5c56b27b7384767 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#779>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEEEYdHEAM9-MLYRtKv6e8zs15CEjFrks5u-8-CgaJpZM4ZmCBC> .

[remcomokveld]

Yeah, I realize that, but an additional artifact could still work right? Or would that involve too much change because this is a Maven project?

[JakeWharton]

It would work, but we might as well do an annotation processor then so that the logic can be shared by both.

…

On Tue, Jan 1, 2019, 4:27 PM Remco Mokveld ***@***.*** wrote: Yeah, I realize that, but an additional artifact could still work right? Or would that involve too much change because this is a Maven project? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#779 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEEEW7BrZ_oS4ojf9TsKG1uO-G4I7Onks5u-9KzgaJpZM4ZmCBC> .