Customize Run / Debug with Command Line Arguments

[JP-Ellis]

Having the ability to quickly run and/or debug an individual test has been absolutely fantastic and in the case of tests and benchmarks, this is very straightforward. It would be nice though if the run/debug could be customized in some way within examples/ (and I'm guessing bin/ too?) where the main function could reasonably expect command line arguments.

I'm not sure what might be the best way to approach this. I can see a two main options:

1. Adding 'Run with args' and 'Debug with args' to the list of possibilities. Four different run/debug options could become a big cluttered, but as this would not apply to most tests / benchmarks, I'm not sure this would be a big issue.
2. Adding a 'Customize' option which would create a pre-filled entry in launch.json allowing for the task to be run this way (and customized). Perhaps this could apply more generally?

[matklad]

Yeah, this is something I very much want to have, but there's a large design space here. The upstream issue is tracked at https://github.com/microsoft/language-server-protocol/issues/944 but, historically, waiting for the upstream (lsp/vscode) wasn't productive.

So perhaps it makes sense to just do it ourselves. I think this mostly requires coding the VS Code side of things. I think the missing pieces are:

• adding a concept of ID to a run configuration, so that the editor can match run config returned by the server to persisted settings for this specific configuration
• adding configuration persistance. They probably want to be saved in workspace configuration file (not sure about this)
• adding templating/prototypical configuration. We should allow tweaking the arguments when running a partciualr test, together with tweaking default arguments when running any tests.
• some UI affordances to move you from a #[test] function in the source code to the corresponding bit of JSON in the config.

[vsrs]

2. Adding a 'Customize' option which would create a pre-filled entry in launch.json allowing for the task to be run this way (and customized). Perhaps this could apply more generally?

For Debug you can use "rust-analyzer.newDebugConfig" command ("Generate launch configuration" in UI). It works exactly like described: creates a new configuration in launch.json allowing to specify args, environment, etc.

Unfortunately it does not work for Run