Anh-Dung Phan

@jtmueller I really appreciate if you could provide a minimal repro. If the internal state is corrupted, it will mark all open statements as unused. We should definitely fix that.

@enricosada Sure. Please do so.

On this particular issue, the explanation is that we haven't handled Abstract attribute yet. We should handle it somewhere around https://github.com/fsprojects/VisualFSharpPowerTools/blob/7af929ea59b6c610a1f0ac0179ed8f40e30ba5f2/src/FSharp.Editing/CodeGeneration/SignatureGenerator.fs#L492-L509 It would be great to modify an existing test...

Sorry for the late reply (I have been travelling in the last few days). A few points: - We should not go into reflection route unless there is no alternative....

Now I think it's best to send a PR to FSharp.Compiler.Service. We still need to add a test to FCS to check whether this works. > I'm currently trying to...

Looking at the test results, it seems like there are internal breaking changes in AST in FCS. Once a new FCS NuGet package is published, we should upgrade and adapt...

@TeaDrivenDev I don't know how you did it. If you take the new FCS and compile the whole VFPT solution against it, there will be compiler errors due to breaking...

@scottf It has been done by Dependabot at https://github.com/nats-io/stan.net/pull/203? Is there any chance to merge that PR and release a new version of STAN.Client NuGet package?

@scottf I upgraded Google.Protobuf.Tools to 3.15.0 in https://github.com/nats-io/stan.net/pull/205 for consistency. We are not ready to move to 3.21.1 yet, sorry.

@matthid I'm pretty sure this is doable by current state of FCS (at least for F#). The issue might be that FSharp.Formatting used an old metadata reader and hasn't been...