fix(iota-move): fix os error 2 for iota move new command using <NAME> with uppercase letter(s)

[roman1e2f5p8s]

Description of change

• Use the provided <NAME> instead of its lowercase variant to obtain the correct and existing path for writing.
• Update <NAME> validation error message: since <NAME> is validated as an Identifier, a valid <NAME> can also start with an uppercase letter or underscore.

Links to any relevant issues

Fixes #4895.

Type of change

• Bug fix (a non-breaking change which fixes an issue)

How the change has been tested

(EDIT: On Linux) Run the following before and after the change:

cargo build
./target/debug/iota move new Xxx

Change checklist

• I have followed the contribution guidelines for this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• I have checked that new and existing unit tests pass locally with my changes

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
apps-backend	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 31, 2025 3:45pm
apps-ui-kit	✅ Ready (Inspect)	Visit Preview		Jan 31, 2025 3:45pm
rebased-explorer	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 31, 2025 3:45pm
wallet-dashboard	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 31, 2025 3:45pm

[lzpap]

Is there any other place in the stack where package names are assumed to be lowercase? @valeriyr
Afaik the on-chain packages are always denoted via their IDs, so the package name is simply a client side concept right now, correct?

[valeriyr]

Is there any other place in the stack where package names are assumed to be lowercase? @valeriyr Afaik the on-chain packages are always denoted via their IDs, so the package name is simply a client-side concept right now, correct?

It looks like the issue can be reproduced only on Linux. Yesterday I tried it locally without changes from this PR on macOS and the result was the following:

./target/debug/iota move new Xxx

tree ./Xxx
./Xxx
├── Move.toml
├── sources
│   └── xxx.move
└── tests
    └── xxx_tests.move

3 directories, 3 files

Xxx is used only like a folder name and as the package name is the toml file name = "Xxx". In external smart contracts, a package is accessible using a named address which is xxx = "0x0" in the example. I think we need to check the changes from this PR carefully before merging.

[valeriyr]

In a Move.lock file we also can find a package name as a reference:

[[move.package]]
name = "Iota"
source = { git = "https://github.com/iotaledger/iota.git", rev = "testnet", subdir = "crates/iota-framework/packages/iota-framework" }

Iota is capitalized, btw.

[roman1e2f5p8s]

@valeriyr, yes, I discovered the issue and tested the changes on Linux, I didn't have the chance to check it on macOS. Nevertheless, I am surprised how and why the issue is not reproduced on macOS: when using ./target/debug/iota move new Xxx, the following line constructs path as ./xxx instead of ./Xxx:

iota/crates/iota-move/src/new.rs

Line 30 in 6179d35

let p = path.unwrap_or_else(|| Path::new(&name));

However, ./xxx might not exist, which results in attempts to write to a non-existing path:

iota/crates/iota-move/src/new.rs

Lines 31 to 44 in 6179d35

	let mut w = std::fs::File::create(
	p.join(SourcePackageLayout::Sources.path())
	.join(format!("{name}.move")),
	)?;
	writeln!(
	w,
	r#"/*
	/// Module: {name}
	module {name}::{name} {{
	}}
	*/"#,
	name = name
	)?;

Also, this

iota/crates/iota-move/src/new.rs

Lines 28 to 29 in 6179d35

	self.new
	.execute(path, [(IOTA_PKG_NAME, IOTA_PKG_PATH)], [(name, "0x0")], "")?;

creates and writes to ./Xxx (I think that's correct), but everything below that will attempt to write to ./xxx.

[alexsporn]

FYI: MacOS is by default case-insensitive

[roman1e2f5p8s]

Thanks, @alexsporn, that explains why the issue is not reproducible on macOS.