How does this differ from BP_NATIVE_IMAGE_BUILD_ARGUMENTS_FILE? We added BP_NATIVE_IMAGE_BUILD_ARGUMENTS_FILE for Quarkus support which does something similar, when it builds it generates a file full of native-image arguments that need to be included in the build command. My worry here is that we have two similar but different features which is confusing for users.

Also, what exactly is an @argfile. I'm not familiar with that. How is native-image handling that argument?

The BP_NATIVE_IMAGE_BUILD_ARGUMENTS_FILE option in the buildpack basically just reads a file, the contents of which are then appended to the arguments list passed to native-image. So the file is just a single line of arguments that will be used by native-image.

How does this differ from BP_NATIVE_IMAGE_BUILD_ARGUMENTS_FILE

The main difference is that we want a way to provide arguments as something that's packaged inside the jar and applied by convention. As I understand it, the BP_NATIVE_IMAGE_BUILD_ARGUMENTS_FILE is an environment variable that the user would still need to configure. We're looking for a way to remove any custom configuration that the user needs to remember.

Also, what exactly is an @argfile. I'm not familiar with that. How is native-image handling that argument?

They're a standard adopted by java command line tools. Some documentation is available here.

There's also a brief comment about them if you run native-image --help:

GraalVM Native Image (https://www.graalvm.org/native-image/)

This tool can ahead-of-time compile Java code to native executables.

...

where options include:

    @argument files       one or more argument files containing options

Closing in favour of #204