feat: accept ETH_RPC_URL env as fork-url alias

[CodeSandwich]

Motivation

Currently commands like forge script don't read the ETH_RPC_URL env variable, you need to to forge script --rpc-url "$ETH_RPC_URL" (or --fork-url or -f, --rpc-url is just one of the aliases). This is unlike all other commands like forge create, which accept ETH_RPC_URL as the --rpc-url alternative.

Solution

Add support for ETH_RPC_URL wherever --fork-url is accepted.

[CodeSandwich]

The integration tests are a rabbit hole to fix. This PR isn't as simple as I was expecting and currently I don't have time to fix everything, so I'll return to it in a few days.

[klkvr]

related: #7755

[yash-atreya]

@CodeSandwich thanks for your contribution. Do you still have the bandwidth to work on this?

[CodeSandwich]

@yash-atreya Thanks for the reminder, TBH I forgot about this PR. I think that I may be able to take another stab this weekend unless somebody wants to take it over.

[CodeSandwich]

@yash-atreya I'm giving up on this PR, feel free to pass it to somebody else or close it for good.

I'm simply unable to work with the integration tests. Running them on my local machine with (make test-foundry with --no-fail-fast added) yields dozens if not hundreds of failing tests, most of which seem to be working on the CI. It smells like an environment mismatch but I have no idea how to prepare my machine to properly run the tests, I haven't found any documentation for it. OTOH the CI seems to be hard failing on tests that are reliably passing on my machine. There's a ton of flaky tests that are adding a lot of noise and I'm not sure if the hard failing ones are actually indicating broken cases or bad luck and flaking 3 times in a row.

I wish all the best to Foundry, I like this tool very much and I want it to succeed. I think that the developer experience may be harming the project, for example I as of now can't see myself contributing to its codebase and probably other outside contributors are facing a similar barrier. I hope that at least the core contributors manage to achieve a good experience and aren't slowed down by issues around the test suite.

[yash-atreya]

@yash-atreya I'm giving up on this PR, feel free to pass it to somebody else or close it for good.

I'm simply unable to work with the integration tests. Running them on my local machine with (make test-foundry with --no-fail-fast added) yields dozens if not hundreds of failing tests, most of which seem to be working on the CI. It smells like an environment mismatch but I have no idea how to prepare my machine to properly run the tests, I haven't found any documentation for it. OTOH the CI seems to be hard failing on tests that are reliably passing on my machine. There's a ton of flaky tests that are adding a lot of noise and I'm not sure if the hard failing ones are actually indicating broken cases or bad luck and flaking 3 times in a row.

I wish all the best to Foundry, I like this tool very much and I want it to succeed. I think that the developer experience may be harming the project, for example I as of now can't see myself contributing to its codebase and probably other outside contributors are facing a similar barrier. I hope that at least the core contributors manage to achieve a good experience and aren't slowed down by issues around the test suite.

Thank you for the contribution and much-needed feedback on local testing for contributors. Will keep this in mind going forward.

[DaniPopes]

It looks like ETH_RPC_URL env var has priority over the CLI flag, cc @mattsse

Tests now pass because we don't need to use this env var anymore

[mattsse]

makes sense but this could be breaking for some setups