feat: Implement print method

[grasshopper47]

Description

Problem*

Resolves #3575

Summary*

Adds print method alongside println to ease printing in special use-cases, such as elements of slices.
Updates appropriate tests, and logging.md documentation.
Add a subchapter to 03_string.md documentation for the fmtstr type

Additional Context

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [Exceptional Case] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[grasshopper47]

@TomAFrench @vezenovm review please

[vezenovm]

Looks good. Could you also add a couple print tests to nargo_cli/tests/execution_success/strings?

[jfecher]

If we add a print function, I think we should reimplement println in terms of that so that we only have 1 primitive instead of two. println could be rewritten as

unconstrained pub fn println<T>(input: T) {
    print(input);
    print("\n");
}

[grasshopper47]

If we add a print function, I think we should reimplement println in terms of that so that we only have 1 primitive instead of two. println could be rewritten as

unconstrained pub fn println<T>(input: T) {
    print(input);
    print("\n");
}

Would

print(f"{input}
");

be a quicker alternative than 2 calls?

[jfecher]

Would [snippet] be a quicker alternative than 2 calls?

Possibly, I'm not quite sure. Feel free to go with whichever approach is easier / more straightforward.

[TomAFrench]

I think this would be preferable to avoid multiple foreign calls being performed.

[grasshopper47]

I think this would be preferable to avoid multiple foreign calls being performed.

It creates a problem: what if you pass a fmtstr down to println, and it passes that down to print ... formatted format strings cannot be printed:

I get the crash above when I run a personal test Noir project through the newly built compiler

I'm surprised this isn't caught at test-time....

As specified in the docs, I ran nix flake check -L, thinking this is how to run the tests. There are countless std::println(f" statements in the .nr tests, why didn't I get an error during testing?

[grasshopper47]

I think this would be preferable to avoid multiple foreign calls being performed.

Pending a reply on the previous point, I have another suggestion:

@TomAFrench Let me know if this is the right idea

[jfecher]

@grasshopper47 format strings should be able to be printed. If I had to guess what was happening, when you switched println to print, we're missing an explicit check for Println to handle format strings before it is actually called. That check, along with any others, would need to be changed to check for Print instead. Alternatively, I'm fine with the oracle change as well (we'll just need to test it to make sure it still works for tuples). Although we should name the intermediate value to make it more obvious what it represents E.g. print_newline.

[grasshopper47]

@grasshopper47 format strings should be able to be printed. If I had to guess what was happening, when you switched println to print, we're missing an explicit check for Println to handle format strings before it is actually called. That check, along with any others, would need to be changed to check for Print instead. Alternatively, I'm fine with the oracle change as well (we'll just need to test it to make sure it still works for tuples). Although we should name the intermediate value to make it more obvious what it represents E.g. print_newline.

Format strings ARE printable, just someplace else, and duplicating that in execute_print would be (needless) extra work IMO. _with_newline param looks "cleaner".

For tuples, my guess is that we'll need to switch the params around, the bool being first, maybe.
I need to understand how to run the tests in order to check, can you help out?

[jfecher]

I need to understand how to run the tests in order to check, can you help out?

You should be able to run them all with cargo test locally

[grasshopper47]

I need to understand how to run the tests in order to check, can you help out?

You should be able to run them all with cargo test locally

Unfortunately it fails to work, hence the question. Look-see:

[grasshopper47]

@TomAFrench @jfecher

I've been trying to run the tests locally to see why 2 of the checks are failing, but I'm unable to do so, see previous comment.

Currently I'm blocked on this.

I've tried:

#[oracle(print)]
unconstrained fn print_oracle<T>(_input: T)

unconstrained pub fn print<T>(input: T) {
    print_oracle(input);
}
unconstrained pub fn println<T>(input: T) {
print(f"{input}
");

This will not work when passing fmtstr to println, as it would lead to a doubly-formatted string. Extracting that is needless extra work.

2nd thing I've tried:

#[oracle(print)]
unconstrained fn print_oracle<T>(_input: T, _with_newline: bool) {}

unconstrained pub fn print<T>(input: T) {
    print_oracle(input, false);
}

unconstrained pub fn println<T>(input: T) {
    print_oracle(input, true);
}

This works fine, except for the failing checks.

If there's no help further possible on this, then I suggest we revert to

unconstrained pub fn println<T>(input: T) {
    print(input);
    print("\n");
}

and call it a day

[jfecher]

@grasshopper47 it runs past that point for me locally. Do you have clang, the c++ stdlib, and libomp installed? If you have nix/direnv I would have thought that'd be handled for you.

[grasshopper47]

@grasshopper47 it runs past that point for me locally. Do you have clang, the c++ stdlib, and libomp installed? If you have nix/direnv I would have thought that'd be handled for you.

Clang install did the trick, many thanks!

[grasshopper47]

@jfecher Something wonky about this branch, I remade the PR.

It contains all the changes requested here.