fixes #171

Hi @eed3si9n
I would appreaciate if you could take a look at this PR. Thank you.

I had the same thought (that the default should be not to generate these imports). However, I was afraid to break somebody's build when they upgraded the sbt-buildinfo plugin.

But yes, instead of what I have done now, it could be:

• Don't generate import scala.Predef._ (and import scala.Any in the ScalaCaseClassRenderer) by default.
• If setting option BuildInfoOption.ExplicitScalaImports it will put these import into the generated code.

Let me know, if you think that it is better to flip it around.
And in that case, what should the name of the option then be? ExplicitScalaImports?

If something fails at build-time, then I think it's ok (as opposed to runtime).
It's already failing someone's build so it's really a matter of which trolley victim group would be larger, and my guess is that very small % of the ppl use -Yno-imports.

Let me know, if you think that it is better to flip it around.

So yeah. Let's do it.

And in that case, what should the name of the option then be? ExplicitScalaImports?

ImportScalaPredef?

Ok, I'll update the pull request tomorrow.

Regarding the name of the option: ScalaCaseClassRenderer generates these imports:

import scala.Predef._
import scala.Any

but I think I can get rid of the scala.Any import by fully qualifying the explicit reference to Any in the generated code.

In that case, I like your suggestion ImportScalaPredef.

It looks like almost all types are fully qualified in the generated code so there is also the possibility that it would be possible to change the code a little so it will compile with and without -Yno-imports without the need for any import of scala.Predef._ at all. That would be the best solution. I'll see if that is possible.

Ok, there are SO many things missing if java.lang and scala.Predef._ is not imported. So to avoid making a lot of changes and risk breaking things, I will go for what we talked about:

• Fully qualify scala.Any, so that is not needed as an explicit import.
• New option ImportScalaPredef that imports just scala.Predef._ for the few that wants to compile with -Yno-imports.

I will leave all the currently, fully qualified types (scala.Int, scala.Boolean, etc.) untouched to keep the changes minimal.

@eed3si9n I have now updated the PR. I have squashed all changes into one commit.
Summary:

• The import scala.Predef._ is no longer generated.
• I have removed the import of scala.Any from ScalaCaseClassRenderer and instead changed the code-generation to emit scala.Any in the only location, where it was necessary.
• I have added the new option BuildInfoOption.ImportScalaPredef that – if present – reintroduces the import scala.Predef._ into the generated code.
• All existing tests have been updated with option BuildInfoOption.ImportScalaPredef so they work exactly as before.
• Two new tests have been added:
  • skipimportscaseobject : Copy of test options but without option BuildInfoOption.ImportScalaPredef and without Scala compiler option -Yno-imports in order to test code generation without the import of scala.Predef._.
  • skipimportscaseclass: Copy of test caseclassrenderer but with same changes as above to test code generation without the import of scala.Predef._.
• Documentation has been updated
  • Description of the new option BuildInfoOption.ImportScalaPredef
  • The line import scala.Predef._ from the sample code in the beginning has been removed as this is no longer generated by default.

Let me hear if it makes sense. :)

@eed3si9n I can see the PR is not linked to issue #171. Not sure how to do that.