Introduce universal compilation host to execution framework

[erak]

Description

This PR implements a universal compiler host for running tests with various compilers (etc. the internal compiler aka. CompilerStack or standard JSON compiler).

Details

It introduces some common types for the compiler inputs and its output. A compilation module that wraps the linked / internal compiler was added. It can be used in a newly added universal compiler host. This host is then used inside the test execution framework.

Outlook

The input / output types introduced by this PR were inspired by the SolidityCompilationFramework. In a future PR the universal compiler host could be used there as well.

Ultimately, this PR makes it possible to implement a compilation module for the standard JSON compiler as well and use that in the pipeline.

[github-actions]

Thank you for your contribution to the Solidity compiler! A team member will follow up shortly.

If you haven't read our contributing guidelines and our review checklist before, please do it now, this makes the reviewing process and accepting your contribution smoother.

If you have any questions or need our help, feel free to post them in the PR or talk to us directly on the #solidity-dev channel on Matrix.

[stackenbotten]

There was an error when running chk_coding_style for commit f79fbaed6453a28a25f6d1cd338074525c4359cf:

Error: Trailing whitespace found:
test/libsolidity/util/compiler/InternalCompiler.cpp:93:        
test/libsolidity/util/compiler/InternalCompiler.cpp:94:    
test/libsolidity/util/compiler/InternalCompiler.cpp:95:    
test/libsolidity/util/compiler/InternalCompiler.cpp:98:    //     std::cout << "sourceName: " << sourceName << std::endl; 
test/libsolidity/util/compiler/InternalCompiler.cpp:104:            

Please check that your changes are working as intended.

[clonker]

Some things need to be ironed out still, on a small scale you should make a pass over the PR and fix all the autos to use auto const& and auto const* where appropriate. Or respective mutable reference/pointer types. Or nothing, if you explicitly want to make a copy :)
Also (but optionally), I personally like to make as much const as possible. For ex., if you have a bool shouldDoSomething = m_compiler.shouldDoSomething() will the sole purpose to pass it on and not modify it further, I declare it as bool const shouldDoSomething = ....
On the bigger scale of things, I am not a big fan of these optionals everywhere to capture experimental solidity in which some data members aren't available. It leads to a lot of bloat. Perhaps a variant with two clean distinct structs would be better. But not sure, really.

[erak]

Hey @clonker, I'm done reworking this. Would you mind having another look?

[clonker]

Some more comments :)

[clonker]

Suggested change

	solAssert(m_sources.find(_sourceName) != m_sources.end(), "No source found.");
	auto const& source = m_sources.find(_sourceName);
	auto const& source = m_sources.find(_sourceName);
	solAssert(source != m_sources.end(), "No source found.");

[clonker]

Suggested change

	m_compilerInput.optimise = solidity::test::CommonOptions::get().optimize;
	m_compilerInput.optimise = CommonOptions::get().optimize;

[clonker]

Suggested change

	auto object = contract->object;
	auto runtimeObject = contract->runtimeObject;
	auto assemblyItems = contract->assemblyItems.value();
	bytes const& object = contract->object;
	bytes const& runtimeObject = contract->runtimeObject;
	auto const& assemblyItems = contract->assemblyItems.value();

[clonker]

Suggested change

	PathGasMeter meter(assemblyItems, solidity::test::CommonOptions::get().evmVersion());
	PathGasMeter meter(assemblyItems, CommonOptions::get().evmVersion());

[clonker]

Suggested change

	auto runtimeAssemblyItems = contract->runtimeAssemblyItems.value();
	auto const& runtimeAssemblyItems = contract->runtimeAssemblyItems.value();

[clonker]

This stuff can be simplified when experimental solidity is skipped

[clonker]

I'd cram that directly into the errorInformation.append. Then you're not prone of making another copy here. It's almost certainly being optimized away but yeah, not entirely certain either :)

[clonker]

unused in the header

[clonker]

Suggested change

	* Abstraction of the internal compiler aka. `CompilerStack`.
	* Abstraction of the `CompilerStack`.

abstraction of the internal internal compiler stack :D

[clonker]

I am not sure if we were over this before, but why is this even stateful, especially since the whole stack gets reset each call to compile?