blackdrag commented on PR #3235:
URL: https://github.com/apache/groovy/pull/3235#issuecomment-6925827749

   @daniellansun Instead of giving a detail review I think we need first to 
talk about the basic problem I see in this PR
   
   I think two aspects of the current implementation need reconsideration:
   
   1. `ModuleInfoNode` is a first-class language construct, but it is currently 
attached to a synthetic `ClassNode` through node metadata. You say "The module 
declaration is not a class (JLS ยง7.3, ยง7.7), and the per-class pipeline is what 
we needed to reuse. " But this makes the module declaration look like auxiliary 
information about a class, which it isn't. It also introduces an artificial 
dependency on the class-node representation and risks the declaration being 
lost or mishandled by compiler transformations. The module declaration should 
be represented explicitly in the compilation-unit model, rather than relying on 
class metadata as its primary representation. At least if this is supposed to 
be a "clean" solution. Of course I see how this is a much bigger change in 
total.
   
   2.  `CompileUnit` should not need to know which class files were produced by 
an external Java compiler. This information should remain within the 
joint-compilation implementation or be exposed through an appropriate 
compilation context.
   
   There are potentially more issues, but changes to these two would complete 
change the PR I assume.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to