LDM Notes for February 7th, 2024

[333fred]

https://github.com/dotnet/csharplang/blob/main/meetings/2024/LDM-2024-02-07.md

Agenda

• Partial type inference
• Breaking change warnings

With this, we are caught back up until the next time I fall behind again. Thanks for your patience while I got all the notes from the past few weeks out.

[yaakov-h]

What was the suggestion for changing <LangVersion>Latest</LangVersion>?

[jnm2]

From #7918:

Alternative: We could "park" the meaning of latest at C# 12, either with or without a warning to say that latest no longer works as you'd expect. This way, users would still get breaking change warnings, and would quickly be discouraged from using latest.

[KennethHoff]

It was a way to go around the fact that "latest" means the newest officially released that's supported by the SDK. If there renew l ends up being breaking changes to the language then simply installing a new SDK could potentially break your code - without even touching it!

[yaakov-h]

That's interesting. I've found Latest to be useful when multitargeting to older runtimes, but still use global.json so that installing a new SDK doesn't immediately upend the whole universe. Without that, yes it does sound quite problematic.

[jnm2]

in case they did not migrate using whatever garden-path approach we create

I'm familiar with https://en.wikipedia.org/wiki/Garden-path_sentence (luring into an initial incorrect interpretation) and https://www.merriam-webster.com/dictionary/lead%20%28someone%29%20down%2Fup%20the%20garden%20path (deceiving), neither of which sound like a lovely approach to create 😄