Move e2e tests from a dedicated crate to tests/ at the workspace root

[agryaznov]

A small enhancement regarding integration tests:

1. We declare a "nominal" timechain package in the workspace Cargo.toml.
2. This allows us to get rid of a dedicated crate for e2e tests, and have them living in tests/ directory, which is neat and in line with general convention for integration tests.

[agryaznov]

@penumbra23 let's come up with the CI settings reliable enough to be accepted for all the future PRs

[dvc94ch]

don't forget to remove the e2e-tests folder...

[agryaznov]

@penumbra23 could port mappings break things for other ci jobs? If so we can rm it for now and disable the e2e flow until that's solved

[haider-rs]

i guess both chain-2-evm and chain-3-evm is mapped to same port. If we change one to something else it shouldnt be an issue.

[agryaznov]

i guess both chain-2-evm and chain-3-evm is mapped to same port. If we change one to something else it shouldnt be an issue.

hmm to me it looks like they're mapped to different ones: 8545 and 8546
(otherwise it wouldn't work locally, but it works)

[haider-rs]

correct my bad.

[agryaznov]

For now let's merge this w the new e2e flow off, and without port mappings, until we figure out proper CI setting on the runners for that with @penumbra23 (tbd in a follow-up PR)

[penumbra23]

One thing that can be done is export the port but avoid a fixed one:

ports:
- 9944

In this way the OS will assign a free one each time.

[agryaznov]

One thing that can be done is export the port but avoid a fixed one:

ports:
- 9944

In this way the OS will assign a free one each time.

this particular one has been there for a quite a while (was not touched in recent PRs) so I keep it as is here

[FlorianFranzen]

@agryaznov This actually added an invalid workflow file still failing CI in some status messages. Just though to let you know as it took me a while to even see where and why the CI was still failing. Nothing to do from your side, I will add a fix #1573.