Create wasmtime::Config from toml

[ludfjig]

This PR makes it possible to create a wasmtime::Config from a toml file and is a first step toward resolving #8784. Overall, I think the path forward is as follows:

1. This PR: Enables loading wasmtime::Config from a toml file, by making CommonOptions derive Deserialize. For example, opt_level is deserialized from values 0, 1, 2, "s" (same as cli flags) instead of "None", "Speed", "SpeedAndSize". Same applies to regalloc_algorithm, compiler and collector. Feedback appreciated on this.
2. Combine cache-config toml with this new toml.
3. Have one source of truth for cli flags and config options.

Feedback greatly appreciated, especially regarding how to deal with nn_graph, as well config_var and keyvalue_in_memory_data, which are all marked serde(skip) for now

[github-actions]

Label Messager: wasmtime:config

It looks like you are changing Wasmtime's configuration options. Make sure to
complete this check list:

• If you added a new Config method, you wrote extensive documentation for
  it.

  Our documentation should be of the following form:

  Short, simple summary sentence.
  
  More details. These details can be multiple paragraphs. There should be
  information about not just the method, but its parameters and results as
  well.
  
  Is this method fallible? If so, when can it return an error?
  
  Can this method panic? If so, when does it panic?
  
  # Example
  
  Optional example here.
  

• If you added a new Config method, or modified an existing one, you
  ensured that this configuration is exercised by the fuzz targets.

  For example, if you expose a new strategy for allocating the next instance
  slot inside the pooling allocator, you should ensure that at least one of our
  fuzz targets exercises that new strategy.

  Often, all that is required of you is to ensure that there is a knob for this
  configuration option in wasmtime_fuzzing::Config (or one
  of its nested structs).

  Rarely, this may require authoring a new fuzz target to specifically test this
  configuration. See our docs on fuzzing for more details.

• If you are enabling a configuration option by default, make sure that it
  has been fuzzed for at least two weeks before turning it on by default.

---

To modify this label's message, edit the .github/label-messager/wasmtime-config.md file.

To add new label messages or remove existing label messages, edit the
.github/label-messager.json configuration file.

Learn more.

[alexcrichton]

Thanks for this! I'm liking how this is shaping up pretty cleanly so far. Would you be up for integrating this into a CLI flag itself as well? For example wasmtime --config cfg.toml or something like that

[ludfjig]

Thanks for this! I'm liking how this is shaping up pretty cleanly so far. Would you be up for integrating this into a CLI flag itself as well? For example wasmtime --config cfg.toml or something like that

Thanks for looking at this. I added a --config flag. Any suggestions on how to best test this?

[alexcrichton]

For tests I'd recommend tests/all/cli_tests.rs or somewhere around there for making a tempdir with a config file

[alexcrichton]

This is looking good to me, was there something else you wanted to address before landing though?

[ludfjig]

Sorry for the delay @alexcrichton. I was wondering if I need to update any documentation, and if the current tests are sufficient. If not, I am ready for merge/reviews.

[alexcrichton]

I think this is good to land, so I'm going to go ahead and approve-and-merge.

If you'd like to help write some docs I suspect that https://github.com/bytecodealliance/wasmtime/blob/main/docs/cli-options.md might be a good location to put this perhaps?

[ludfjig]

@alexcrichton I had some problems getting the test suite with the i686 target to pass, but it should hopefully be fixed now by using wasm-max-stack=0 which is explicitly disallowed and will cause an error.

[alexcrichton]

Let's see if that's spurious...

[alexcrichton]

Ah sorry it looks like #10123 landed which causes conflicts with this. I'm happy to help out with resolving conflicts if needed.

[ludfjig]

@alexcrichton no worries. The only question I have is about the fmt::Display impl for CommonOptions. If configured using a --config <path>, then the output will potentially be incorrect, since it will look like all options were provided on cli instead of from a toml. Is this acceptable?

[alexcrichton]

I think that's reasonable yeah, the Display implementation here is sort of best-effort so printing out --config $path for when the field is Some I think is good enough