feat: dfx start --emulator

[nomeata]

mostly cargo-culting, but it seems to work:

~/dfinity/sdk/test $ cargo run --  start --emulator
warning: /home/jojo/dfinity/sdk/src/dfx/Cargo.toml: dependency (ic-agent) specification is ambiguous. Only one of `branch`, `tag` or `rev` is allowed. This will be considered an error in future versions
warning: /home/jojo/dfinity/sdk/src/dfx/Cargo.toml: dependency (ic-identity-hsm) specification is ambiguous. Only one of `branch`, `tag` or `rev` is allowed. This will be considered an error in future versions
warning: /home/jojo/dfinity/sdk/src/dfx/Cargo.toml: dependency (ic-types) specification is ambiguous. Only one of `branch`, `tag` or `rev` is allowed. This will be considered an error in future versions
warning: /home/jojo/dfinity/sdk/src/dfx/Cargo.toml: dependency (ic-utils) specification is ambiguous. Only one of `branch`, `tag` or `rev` is allowed. This will be considered an error in future versions
    Finished dev [unoptimized + debuginfo] target(s) in 0.14s
     Running `/home/jojo/dfinity/sdk/target/x86_64-unknown-linux-musl/debug/dfx start --emulator`
Starting ic-ref...
Running at http://127.0.0.1:8001/
Starting webserver on port 8001 for replica at "http://localhost:8001"
binding to: V4(127.0.0.1:8000)
replica(s): http://localhost:8001/

^CStopping ic-ref...
Stopped.

[nomeata]

Probably someone needs to re-do most of it with good rust/actix/general software architecture. So far I tried to touch as little code as possible.

Probably at least the signal communication between the replica_webserver_coordinator and the replica/emulator actor should be unified (the replica_webserver_coordinator probably doesn’t care which backend is used) done

Also, lots of test mentions replica… maybe replace that with a more general term that simply applies to either.

[nomeata]

Ok, ic-ref e2e tests now also run with dfx start

[nomeata]

This is probably in a reviewable form. Or, maybe even better, in a take-overable form

[p-shahi]

You probably want @ericswanson-dfinity or @hansl to review this. I added support for the emulator flag to dfx replica command.

Also, lots of test mentions replica… maybe replace that with a more general term that simply applies to either.

#1373 Hope this addresses that concern.

[nomeata]

Yeah, maybe I’ll leave it to @hansl to press the button; he’ll know better if we want to start exposing this now, or some other time.

[hansl]

Hey. I don't think there should be a big upkeep on this PR and we're all busy with Genesis stuff.

I'll try to squeeze a review in between meetings, but it might take a few days. Thanks

[p-shahi]

@hansl @nomeata this PR has sufficient review/approval and by virtue of the e2e tests passing (they run against dfx start --emulator rather than a raw ic-ref executable) can we just merge this instead of letting it live on and on?

(test failures observed here and there are due to flaky tests)

[nomeata]

I wouldn’t object, it may already be useful just for us to debug things and compare behaviours.

As for the wider Emulator projects, @p-shahi’s observation that ic-ref (or rather, the Wasm interpreter) is too slow to run the wallet canister (or is the wallet canister too expensive?… ;-)) makes me less sure about how quickly we can push this out to developers.

[nomeata]

Whoa, exciting. Let’s see what all breaks