Should we use an enum to encode fields?

[tesuji]

rust/src/librustc_span/lib.rs

Lines 1079 to 1088 in 668a34e

	pub struct SourceFile {
	/// The name of the file that the source came from. Source that doesn't
	/// originate from files has names between angle brackets by convention
	/// (e.g., `<anon>`).
	pub name: FileName,
	/// `true` if the `name` field above has been modified by `--remap-path-prefix`.
	pub name_was_remapped: bool,
	/// The unmapped path of the file that the source came from.
	/// Set to `None` if the `SourceFile` was imported from an external crate.
	pub unmapped_path: Option<FileName>,

enum Name {
    /// The name of the file that the source came from. Source that doesn't 
    /// originate from files has names between angle brackets by convention 
    /// (e.g., `<anon>`). 
    Normal(FileName),
    /// FileName modified by `--remap-path-prefix`. 
    Remapped(FileName),
}

pub struct SourceFile { 
    pub name: Name,
    /// The unmapped path of the file that the source came from. 
    /// Set to `None` if the `SourceFile` was imported from an external crate. 
    pub unmapped_path: Option<FileName>, 
    ...
}

[Nicholas-Baron]

I started writing a naive implementation.

1. Name would have to be pub if the name field in SourceFile is pub
2. Name has to derive the following at a minimum: Clone, Debug, PartialEq, Eq, Encodable, Decodable, Display
3. There are some places in the current code where whether a file was remapped is not taken into account. Should there be something like fn unwrap(&self) -> FileName on Name?

PS: I will put a branch up once everything is compiling, unless someone would like to see the code earlier.

[tesuji]

Nice, could you open a draft PR?

[Nicholas-Baron]

@lzutao Draft PR opened.

[Nicholas-Baron]

@rustbot claim

[matklad]

Super drive by comment, but my gut feeling is that the better fix is to add a type for (name, name_was_remapped, unmpapped_path) triple, rather than go for surface level refactoring of eliminated the bool.

Wild-guessing even further, I'd suggest

pub struct SourceFileName {
    // Note absence of pub
    name: FileName,
    unmapped_name: Option<FileName>
}

impl SourceFileName {
    pub fn name(&self) -> &FileName { &self.name } 
    pub fn unmapped_name(&self) -> &FileName { self.unmpped_name.as_ref().unwrap_or(self.name()) }
}

[petrochenkov]

Yeah, #75616 looks like a regression rather than improvement to me.

[Nicholas-Baron]

@petrochenkov @matklad I have started working on the suggestion. The crate itself and some of its dependents compile.

[Nicholas-Baron]

@matklad Your wild guess seems to be working.

@petrochenkov I hope now #75616 is not a regression.

[Nicholas-Baron]

@rustbot unclaim

I am too distracted to finish the pull request.

[rossmacarthur]

Triage: looks like this was fixed in #83813 :)