Closed ericvergnaud closed 9 months ago
Quickly skimmed through the changes. Since I suppose at some point ANTLR 5 will get rid of the Java runtime, I'd say it's all good as long as tests pass.
Just a couple of commits I'd revisit: e8ae531, 9c125e4
Better not make stuff nullable to avoid all those !!
assertions. I recall in the Strumenta repository I used an ad-hoc object
to replace a null
value.
Quickly skimmed through the changes. Since I suppose at some point ANTLR 5 will get rid of the Java runtime, I'd say it's all good as long as tests pass.
Just a couple of commits I'd revisit: e8ae531, 9c125e4 Better not make stuff nullable to avoid all those
!!
assertions. I recall in the Strumenta repository I used an ad-hocobject
to replace anull
value.
Actually the nullables are required to make the tests pass. That said your suggestions are welcome and we might change that in favor of your proposed strategy. I'd be in favor of doing that in a separate PR such that we know what we're breaking and we can measure the impact on performance ? See #44
@KvanTTT @lppedd All tests pass !!! Your comments are welcome. Changing the title of the PR.
@KvanTTT given the broad scope of this PR, I had to merge your latest changes from dev into this branch. It wasn't seamless, so you might want to double-check the affected files.
@KvanTTT any comments ?
No, I'm on vacation. Feel free to merge without my approval.
This is far from being mergeable (many tests fail) but I thought it might be worth starting to look at the changes and provide comments.