Brief thought: D from Java/.net could be another write barrier benefit

Core D Development Statistics

• 13 bugs fixed
• 15 bugs and enhancement requests opened
• Bugs closed WONTFIX:
  • opOpAssign should be automatically visible for types.
  • Attribute hiding error or warning
  • opDispatch doesn't forward opCall
  • One missing statement is not reachable error message
• Bugs closed INVALID:
  • Templated default constructor not called
  • Can't subtract positive number from length of array of struct with disabled default construction
• 24 pull requests merged into the language: 11 into DMD and druntime and 13 into Phobos
• 0 pull requests merged into the website.

In the community

Community announcements

• Beerconf for dconf online 2022

See more at the announce forum.

Another write barrier benefit

I've previously heard reports of D code being called from C# not working right and there was another report of D code being called from Java not working right on the forum this week. See: https://forum.dlang.org/post/tmigfb$1jmk$1@digitalmars.com

Ali diagnosed it to a thread the Java runtime created without notifying the D runtime. With a write barrier, it'd be possible to obviate the need to notify the D runtime entirely; it would be done lazily at the write barrier boundary. This would probably help these cases too.