Add logd<T>(a: T) as a reporting/printing mechanism

[emilyaherbert]

Per discussion on Slack, it would be nice to have println! and print! for devs 😄

Per @nfurfaro:

Would be nice to integrate this with the tooling, so that println! in a contract would work for debugging on a local node (like hardhat’s console.sol , and would function like the Rust println! in a script.

It was such a win for debugging Solidity when console.sol was introduced.

ideally, forc could strip out these println! s when compilling for release, and leave them in for debugging

no, and that’s why solidity doesn’t have it. Hardhat added a special console.sol contract you could inherit from which added the capability, but only when running against a local hardhat node.

[adlerjohn]

This can largely be reproduced with LOGD

Steps:

• Bump fuel-core in CI to v0.5.0. #1044 added LOGD support for inline asm.
• Add size_of functionality. #818 added size_of, so the stdlib can know how many bytes to LOGD.
• Add an is_reference_type::<T>() builtin. #1121 will add is_reference_type, which is needed for the stdlib to know if it should dereference a type before LOGDing it.

With all that together you can have a generic logd<T>(a: T) function in the stdlib that logs an arbitrary stack type. Then you just print out all the logs in your Rust test.

[emilyaherbert]

@adlerjohn via slack:

println! can largely be replaced by logd, but the SDK needs to support that. We should start with that, and if it's insufficient then implement some sort of standardization around println! so that the SDK knows "oh this LOGD is a print"