Should finish be added to nvim-dap?

[mwcz]

Hi! I'm a heavy user of nvim-dap and nvim-dap-ui, with codelldb for debugging Rust. Rust makes heavy use of implicit return values, like this super contrived example:

/* 1 */  fn my_kinda_rand() -> f32 {
/* 2 */    (1.0 + rand()) / 2.0
/* 3 */  }

My experience with nvim-dap is that it's not possible to put a breakpoint on line 2 and inspect the value. To get around that, I frequently bind the value to a variable, which I can then inspect with the debugger.

/* 1 */  fn my_kinda_rand() -> f32 {
/* 2 */    let r = (1.0 + rand()) / 2.0;
/* 3 */    r
/* 4 */  }

The problem is that this leads to a lot of non-idiomatic Rust, and clippy (Rust's linter) doesn't like it. It's also harder to take this approach when debugging third party code.

Codelldb and gdb both have finish, which runs the current stack frame to completion, and prints the return value.

Perfect! The only hangup is that I haven't found a way to surface that return value in nvim-dap. Is this a feature that others would find useful? Is there already a way to do it that I missed? Thanks for the discussion.

[mfussenegger]

This relates a bit to microsoft/debug-adapter-protocol#458

There's currently nothing in the specification that would allow nvim-dap to surface this in a way that works across debug adapters.

My workaround so far is to just select the whole expression and then use the hover widget on it

[mwcz]

I see, 🤞 the dap proposal is accepted, and thank you for the hover tip, that will help a ton in the meantime.

[mwcz]

It seems that's closed now, but I don't grok what the resolution was. I see mention of a returnValue added after stepOut, which would do the trick I suppose.