compiler/rustc_lint/src/nonstandard_style.rs

use rustc_abi::ExternAbi;
use rustc_ast as ast;
use rustc_attr_parsing::AttributeParser;
use rustc_errors::{Applicability, Diag, DiagCtxtHandle, Diagnostic, Level};
use rustc_hir as hir;
use rustc_hir::attrs::{AttributeKind, ReprAttr};
use rustc_hir::def::{DefKind, Res};
use rustc_hir::def_id::DefId;
use rustc_hir::intravisit::{FnKind, Visitor};
use rustc_hir::{Attribute, GenericParamKind, PatExprKind, PatKind, find_attr};
use rustc_middle::hir::nested_filter::All;
use rustc_middle::ty::AssocContainer;
use rustc_session::config::CrateType;
use rustc_session::{declare_lint, declare_lint_pass};
use rustc_span::def_id::LocalDefId;
use rustc_span::{BytePos, Ident, Span, sym};

use crate::lints::{
    NonCamelCaseType, NonCamelCaseTypeSub, NonSnakeCaseDiag, NonSnakeCaseDiagSub,
    NonUpperCaseGlobal, NonUpperCaseGlobalSub, NonUpperCaseGlobalSubTool,
};
use crate::{EarlyContext, EarlyLintPass, LateContext, LateLintPass, LintContext};

declare_lint! {
    /// The `non_camel_case_types` lint detects types, variants, traits and
    /// type parameters that don't have camel case names.
    ///
    /// ### Example
    ///
    /// ```rust
    /// struct my_struct;
    /// ```
    ///
    /// {{produces}}
    ///
    /// ### Explanation
    ///
    /// The preferred style for these identifiers is to use "camel case", such
    /// as `MyStruct`, where the first letter should not be lowercase, and
    /// should not use underscores between letters. Underscores are allowed at
    /// the beginning and end of the identifier, as well as between
    /// non-letters (such as `X86_64`).
    pub NON_CAMEL_CASE_TYPES,
    Warn,
    "types, variants, traits and type parameters should have camel case names"
}

declare_lint_pass!(NonCamelCaseTypes => [NON_CAMEL_CASE_TYPES]);

/// Some unicode characters *have* case, are considered upper, title, or lower case, but they *can't*
/// be title cased or lower cased. For the purposes of the lint suggestion, we care about being able
/// to change the char's case.
fn char_has_case(c: char) -> bool {
    !c.to_lowercase().eq(c.to_titlecase())
}

/// FIXME: we should add a more efficient version
/// in the stdlib for this
fn changes_when_titlecased(c: char) -> bool {
    !c.to_titlecase().eq([c])
}

// contains a capitalisable character followed by, or preceded by, an underscore,
// or contains an uppercase character that changes when titlecased,
// or contains `__`
fn not_camel_case(s: &str) -> bool {
    let mut last = '\0';
    s.chars().any(|snd| {
        let fst = std::mem::replace(&mut last, snd);
        match (fst, snd) {
            ('_', '_') => return true,
            ('_', _) if char_has_case(snd) => return true,
            (_, '_') if char_has_case(fst) => return true,
            _ => snd.is_uppercase() && changes_when_titlecased(snd),
        }
    })
}

fn is_upper_camel_case(name: &str) -> bool {
    let name = name.trim_matches('_');
    let Some(first) = name.chars().next() else {
        return true;
    };

    // some scripts don't have a concept of upper/lowercase
    !(changes_when_titlecased(first) || not_camel_case(name))
}

fn to_upper_camel_case(s: &str) -> String {
    s.trim_matches('_')
        .split('_')
        .filter(|component| !component.is_empty())
        .map(|component| {
            let mut camel_cased_component = String::new();

            let mut new_word = true;
            let mut prev_is_lower_case = true;
            let mut prev_is_lowercased_sigma = false;

            for c in component.chars() {
                // Preserve the case if an uppercase letter follows a lowercase letter, so that
                // `camelCase` is converted to `CamelCase`.
                if prev_is_lower_case && (c.is_uppercase() | c.is_titlecase()) {
                    new_word = true;
                }

                if new_word {
                    camel_cased_component.extend(c.to_titlecase());
                } else {
                    camel_cased_component.extend(c.to_lowercase());
                }

                prev_is_lower_case = c.is_lowercase() || c.is_titlecase();
                prev_is_lowercased_sigma = !new_word && c == 'Σ';
                new_word = false;
            }

            if prev_is_lowercased_sigma {
                camel_cased_component.pop();
                camel_cased_component.push('ς');
            }

            camel_cased_component
        })
        .fold((String::new(), None), |(acc, prev): (String, Option<String>), next| {
            // separate two components with an underscore if their boundary cannot
            // be distinguished using an uppercase/lowercase case distinction
            let join = if let Some(prev) = prev {
                let l = prev.chars().last().unwrap();
                let f = next.chars().next().unwrap();
                !char_has_case(l) && !char_has_case(f)
            } else {
                false
            };
            (acc + if join { "_" } else { "" } + &next, Some(next))
        })
        .0
}

impl NonCamelCaseTypes {
    fn check_case(&self, cx: &EarlyContext<'_>, sort: &str, ident: &Ident) {
        let name = ident.name.as_str();

        if !is_upper_camel_case(name) {
            let cc = to_upper_camel_case(name);
            let sub = if *name != cc {
                NonCamelCaseTypeSub::Suggestion { span: ident.span, replace: cc }
            } else {
                NonCamelCaseTypeSub::Label { span: ident.span }
            };
            cx.emit_span_lint(
                NON_CAMEL_CASE_TYPES,
                ident.span,
                NonCamelCaseType { sort, name, sub },
            );
        }
    }
}

impl EarlyLintPass for NonCamelCaseTypes {
    fn check_item(&mut self, cx: &EarlyContext<'_>, it: &ast::Item) {
        let has_repr_c = matches!(
            AttributeParser::parse_limited(cx.sess(), &it.attrs, &[sym::repr]),
            Some(Attribute::Parsed(AttributeKind::Repr { reprs, ..})) if reprs.iter().any(|(r, _)| r == &ReprAttr::ReprC)
        );

        if has_repr_c {
            return;
        }

        match &it.kind {
            ast::ItemKind::TyAlias(ast::TyAlias { ident, .. })
            | ast::ItemKind::Enum(ident, ..)
            | ast::ItemKind::Struct(ident, ..)
            | ast::ItemKind::Union(ident, ..) => self.check_case(cx, "type", ident),
            ast::ItemKind::Trait(ast::Trait { ident, .. }) => self.check_case(cx, "trait", ident),
            ast::ItemKind::TraitAlias(ast::TraitAlias { ident, .. }) => {
                self.check_case(cx, "trait alias", ident)
            }

            // N.B. This check is only for inherent associated types, so that we don't lint against
            // trait impls where we should have warned for the trait definition already.
            ast::ItemKind::Impl(ast::Impl { of_trait: None, items, .. }) => {
                for it in items {
                    // FIXME: this doesn't respect `#[allow(..)]` on the item itself.
                    if let ast::AssocItemKind::Type(alias) = &it.kind {
                        self.check_case(cx, "associated type", &alias.ident);
                    }
                }
            }
            _ => (),
        }
    }

    fn check_trait_item(&mut self, cx: &EarlyContext<'_>, it: &ast::AssocItem) {
        if let ast::AssocItemKind::Type(alias) = &it.kind {
            self.check_case(cx, "associated type", &alias.ident);
        }
    }

    fn check_variant(&mut self, cx: &EarlyContext<'_>, v: &ast::Variant) {
        self.check_case(cx, "variant", &v.ident);
    }

    fn check_generic_param(&mut self, cx: &EarlyContext<'_>, param: &ast::GenericParam) {
        if let ast::GenericParamKind::Type { .. } = param.kind {
            self.check_case(cx, "type parameter", &param.ident);
        }
    }
}

declare_lint! {
    /// The `non_snake_case` lint detects variables, methods, functions,
    /// lifetime parameters and modules that don't have snake case names.
    ///
    /// ### Example
    ///
    /// ```rust
    /// let MY_VALUE = 5;
    /// ```
    ///
    /// {{produces}}
    ///
    /// ### Explanation
    ///
    /// The preferred style for these identifiers is to use "snake case",
    /// where all the characters are in lowercase, with words separated with a
    /// single underscore, such as `my_value`.
    pub NON_SNAKE_CASE,
    Warn,
    "variables, methods, functions, lifetime parameters and modules should have snake case names"
}

declare_lint_pass!(NonSnakeCase => [NON_SNAKE_CASE]);

impl NonSnakeCase {
    fn to_snake_case(mut name: &str) -> String {
        let mut words = vec![];
        // Preserve leading underscores
        name = name.trim_start_matches(|c: char| {
            if c == '_' {
                words.push(String::new());
                true
            } else {
                false
            }
        });
        for s in name.split('_') {
            let mut last_upper = false;
            let mut buf = String::new();
            if s.is_empty() {
                continue;
            }
            for ch in s.chars() {
                if !buf.is_empty()
                    && buf != "'"
                    && (ch.is_uppercase() || ch.is_titlecase())
                    && !last_upper
                {
                    // We lowercase only at the end, to handle final sigma correctly
                    words.push(buf.to_lowercase());
                    buf = String::new();
                }
                last_upper = ch.is_uppercase() || ch.is_titlecase();
                buf.push(ch);
            }
            // We lowercase only at the end, to handle final sigma correctly
            words.push(buf.to_lowercase());
        }
        words.join("_")
    }

    /// Checks if a given identifier is snake case, and reports a diagnostic if not.
    fn check_snake_case(&self, cx: &LateContext<'_>, sort: &str, ident: &Ident) {
        fn is_snake_case(ident: &str) -> bool {
            if ident.is_empty() {
                return true;
            }
            let ident = ident.trim_start_matches('\'');
            let ident = ident.trim_matches('_');

            if ident.contains("__") {
                return false;
            }

            // This correctly handles letters in languages with and without
            // cases, as well as numbers and underscores.
            // FIXME: we should add a standard library impl of `c.to_lowercase().eq([c])`
            ident.chars().all(|c| c.to_lowercase().eq([c]))
        }

        let name = ident.name.as_str();

        if !is_snake_case(name) {
            let span = ident.span;
            let sc = NonSnakeCase::to_snake_case(name);
            // We cannot provide meaningful suggestions
            // if the characters are in the category of "Uppercase Letter".
            let sub = if name != sc {
                // We have a valid span in almost all cases, but we don't have one when linting a
                // crate name provided via the command line.
                if !span.is_dummy() {
                    let sc_ident = Ident::from_str_and_span(&sc, span);
                    if sc_ident.is_reserved() {
                        // We shouldn't suggest a reserved identifier to fix non-snake-case
                        // identifiers. Instead, recommend renaming the identifier entirely or, if
                        // permitted, escaping it to create a raw identifier.
                        if sc_ident.name.can_be_raw() {
                            NonSnakeCaseDiagSub::RenameOrConvertSuggestion {
                                span,
                                suggestion: sc_ident,
                            }
                        } else {
                            NonSnakeCaseDiagSub::SuggestionAndNote { sc, span }
                        }
                    } else {
                        NonSnakeCaseDiagSub::ConvertSuggestion { span, suggestion: sc }
                    }
                } else {
                    NonSnakeCaseDiagSub::Help { sc }
                }
            } else {
                NonSnakeCaseDiagSub::Label { span }
            };
            cx.emit_span_lint(NON_SNAKE_CASE, span, NonSnakeCaseDiag { sort, name, sub });
        }
    }
}

impl<'tcx> LateLintPass<'tcx> for NonSnakeCase {
    fn check_mod(&mut self, cx: &LateContext<'_>, _: &'tcx hir::Mod<'tcx>, id: hir::HirId) {
        if id != hir::CRATE_HIR_ID {
            return;
        }

        // Issue #45127: don't enforce `snake_case` for binary crates as binaries are not intended
        // to be distributed and depended on like libraries. The lint is not suppressed for cdylib
        // or staticlib because it's not clear what the desired lint behavior for those are.
        if cx.tcx.crate_types().iter().all(|&crate_type| crate_type == CrateType::Executable) {
            return;
        }

        let crate_ident = if let Some(name) = &cx.tcx.sess.opts.crate_name {
            Some(Ident::from_str(name))
        } else {
            find_attr!(cx.tcx, crate, CrateName{name, name_span,..} => (name, name_span)).map(
                |(&name, &span)| {
                    // Discard the double quotes surrounding the literal.
                    let sp = cx
                        .sess()
                        .source_map()
                        .span_to_snippet(span)
                        .ok()
                        .and_then(|snippet| {
                            let left = snippet.find('"')?;
                            let right = snippet.rfind('"').map(|pos| snippet.len() - pos)?;

                            Some(
                                span.with_lo(span.lo() + BytePos(left as u32 + 1))
                                    .with_hi(span.hi() - BytePos(right as u32)),
                            )
                        })
                        .unwrap_or(span);

                    Ident::new(name, sp)
                },
            )
        };

        if let Some(ident) = &crate_ident {
            self.check_snake_case(cx, "crate", ident);
        }
    }

    fn check_generic_param(&mut self, cx: &LateContext<'_>, param: &hir::GenericParam<'_>) {
        if let GenericParamKind::Lifetime { .. } = param.kind {
            self.check_snake_case(cx, "lifetime", &param.name.ident());
        }
    }

    fn check_fn(
        &mut self,
        cx: &LateContext<'_>,
        fk: FnKind<'_>,
        _: &hir::FnDecl<'_>,
        _: &hir::Body<'_>,
        _: Span,
        id: LocalDefId,
    ) {
        match &fk {
            FnKind::Method(ident, sig, ..) => match cx.tcx.associated_item(id).container {
                AssocContainer::InherentImpl => {
                    if sig.header.abi != ExternAbi::Rust && find_attr!(cx.tcx, id, NoMangle(..)) {
                        return;
                    }
                    self.check_snake_case(cx, "method", ident);
                }
                AssocContainer::Trait => {
                    self.check_snake_case(cx, "trait method", ident);
                }
                AssocContainer::TraitImpl(_) => {}
            },
            FnKind::ItemFn(ident, _, header) => {
                // Skip foreign-ABI #[no_mangle] functions (Issue #31924)
                if header.abi != ExternAbi::Rust && find_attr!(cx.tcx, id, NoMangle(..)) {
                    return;
                }
                self.check_snake_case(cx, "function", ident);
            }
            FnKind::Closure => (),
        }
    }

    fn check_item(&mut self, cx: &LateContext<'_>, it: &hir::Item<'_>) {
        if let hir::ItemKind::Mod(ident, _) = it.kind {
            self.check_snake_case(cx, "module", &ident);
        }
    }

    fn check_ty(&mut self, cx: &LateContext<'_>, ty: &hir::Ty<'_, hir::AmbigArg>) {
        if let hir::TyKind::FnPtr(hir::FnPtrTy { param_idents, .. }) = &ty.kind {
            for param_ident in *param_idents {
                if let Some(param_ident) = param_ident {
                    self.check_snake_case(cx, "variable", param_ident);
                }
            }
        }
    }

    fn check_trait_item(&mut self, cx: &LateContext<'_>, item: &hir::TraitItem<'_>) {
        if let hir::TraitItemKind::Fn(_, hir::TraitFn::Required(param_idents)) = item.kind {
            self.check_snake_case(cx, "trait method", &item.ident);
            for param_ident in param_idents {
                if let Some(param_ident) = param_ident {
                    self.check_snake_case(cx, "variable", param_ident);
                }
            }
        }
    }

    fn check_pat(&mut self, cx: &LateContext<'_>, p: &hir::Pat<'_>) {
        if let PatKind::Binding(_, hid, ident, _) = p.kind {
            if let hir::Node::PatField(field) = cx.tcx.parent_hir_node(hid) {
                if !field.is_shorthand {
                    // Only check if a new name has been introduced, to avoid warning
                    // on both the struct definition and this pattern.
                    self.check_snake_case(cx, "variable", &ident);
                }
                return;
            }
            self.check_snake_case(cx, "variable", &ident);
        }
    }

    fn check_struct_def(&mut self, cx: &LateContext<'_>, s: &hir::VariantData<'_>) {
        for sf in s.fields() {
            self.check_snake_case(cx, "structure field", &sf.ident);
        }
    }
}

declare_lint! {
    /// The `non_upper_case_globals` lint detects static items that don't have
    /// uppercase identifiers.
    ///
    /// ### Example
    ///
    /// ```rust
    /// static max_points: i32 = 5;
    /// ```
    ///
    /// {{produces}}
    ///
    /// ### Explanation
    ///
    /// The preferred style is for static item names to use all uppercase
    /// letters such as `MAX_POINTS`.
    pub NON_UPPER_CASE_GLOBALS,
    Warn,
    "static constants should have uppercase identifiers"
}

declare_lint_pass!(NonUpperCaseGlobals => [NON_UPPER_CASE_GLOBALS]);

struct NonUpperCaseGlobalGenerator<'a, F: FnOnce() -> NonUpperCaseGlobal<'a>> {
    callback: F,
}

impl<'a, 'b, F: FnOnce() -> NonUpperCaseGlobal<'b>> Diagnostic<'a, ()>
    for NonUpperCaseGlobalGenerator<'b, F>
{
    fn into_diag(self, dcx: DiagCtxtHandle<'a>, level: Level) -> Diag<'a, ()> {
        let Self { callback } = self;
        callback().into_diag(dcx, level)
    }
}

impl NonUpperCaseGlobals {
    fn check_upper_case(cx: &LateContext<'_>, sort: &str, did: Option<LocalDefId>, ident: &Ident) {
        let name = ident.name.as_str();
        // FIXME: we should add a more efficient version
        // in the stdlib for `c.to_uppercase().eq([c])`
        if !name.chars().all(|c| c.to_uppercase().eq([c])) {
            let uc = NonSnakeCase::to_snake_case(name).to_uppercase();

            // If the item is exported, suggesting changing its name would be a breaking change
            // and could break users without a "nice" applicable fix, so let's avoid it.
            let can_change_usages = if let Some(did) = did {
                !cx.tcx.effective_visibilities(()).is_exported(did)
            } else {
                false
            };

            // We cannot provide meaningful suggestions
            // if the characters are in the category of "Lowercase Letter".
            let sub = if *name != uc {
                NonUpperCaseGlobalSub::Suggestion {
                    span: ident.span,
                    replace: uc.clone(),
                    applicability: if can_change_usages {
                        Applicability::MachineApplicable
                    } else {
                        Applicability::MaybeIncorrect
                    },
                }
            } else {
                NonUpperCaseGlobalSub::Label { span: ident.span }
            };

            struct UsageCollector<'a, 'tcx> {
                cx: &'tcx LateContext<'a>,
                did: DefId,
                collected: Vec<Span>,
            }

            impl<'v, 'tcx> Visitor<'v> for UsageCollector<'v, 'tcx> {
                type NestedFilter = All;

                fn maybe_tcx(&mut self) -> Self::MaybeTyCtxt {
                    self.cx.tcx
                }

                fn visit_path(
                    &mut self,
                    path: &rustc_hir::Path<'v>,
                    _id: rustc_hir::HirId,
                ) -> Self::Result {
                    if let Some(final_seg) = path.segments.last()
                        && final_seg.res.opt_def_id() == Some(self.did)
                    {
                        self.collected.push(final_seg.ident.span);
                    }
                }
            }

            let callback = || {
                // Compute usages lazily as it can expansive and useless when the lint is allowed.
                // cf. https://github.com/rust-lang/rust/pull/142645#issuecomment-2993024625
                let usages = if can_change_usages
                    && *name != uc
                    && let Some(did) = did
                {
                    let mut usage_collector =
                        UsageCollector { cx, did: did.to_def_id(), collected: Vec::new() };
                    cx.tcx.hir_walk_toplevel_module(&mut usage_collector);
                    usage_collector
                        .collected
                        .into_iter()
                        .map(|span| NonUpperCaseGlobalSubTool { span, replace: uc.clone() })
                        .collect()
                } else {
                    vec![]
                };

                NonUpperCaseGlobal { sort, name, sub, usages }
            };
            cx.emit_span_lint(
                NON_UPPER_CASE_GLOBALS,
                ident.span,
                NonUpperCaseGlobalGenerator { callback },
            );
        }
    }
}

impl<'tcx> LateLintPass<'tcx> for NonUpperCaseGlobals {
    fn check_item(&mut self, cx: &LateContext<'_>, it: &hir::Item<'_>) {
        let attrs = cx.tcx.hir_attrs(it.hir_id());
        match it.kind {
            hir::ItemKind::Static(_, ident, ..) if !find_attr!(attrs, NoMangle(..)) => {
                NonUpperCaseGlobals::check_upper_case(
                    cx,
                    "static variable",
                    Some(it.owner_id.def_id),
                    &ident,
                );
            }
            hir::ItemKind::Const(ident, ..) => {
                NonUpperCaseGlobals::check_upper_case(
                    cx,
                    "constant",
                    Some(it.owner_id.def_id),
                    &ident,
                );
            }
            _ => {}
        }
    }

    fn check_trait_item(&mut self, cx: &LateContext<'_>, ti: &hir::TraitItem<'_>) {
        if let hir::TraitItemKind::Const(..) = ti.kind {
            NonUpperCaseGlobals::check_upper_case(cx, "associated constant", None, &ti.ident);
        }
    }

    fn check_impl_item(&mut self, cx: &LateContext<'_>, ii: &hir::ImplItem<'_>) {
        if let hir::ImplItemKind::Const(..) = ii.kind
            && let hir::ImplItemImplKind::Inherent { .. } = ii.impl_kind
        {
            NonUpperCaseGlobals::check_upper_case(cx, "associated constant", None, &ii.ident);
        }
    }

    fn check_pat(&mut self, cx: &LateContext<'_>, p: &hir::Pat<'_>) {
        // Lint for constants that look like binding identifiers (#7526)
        if let PatKind::Expr(hir::PatExpr {
            kind: PatExprKind::Path(hir::QPath::Resolved(None, path)),
            ..
        }) = p.kind
        {
            if let Res::Def(DefKind::Const { .. }, _) = path.res
                && let [segment] = path.segments
            {
                NonUpperCaseGlobals::check_upper_case(
                    cx,
                    "constant in pattern",
                    None,
                    &segment.ident,
                );
            }
        }
    }

    fn check_generic_param(&mut self, cx: &LateContext<'_>, param: &hir::GenericParam<'_>) {
        if let GenericParamKind::Const { .. } = param.kind {
            NonUpperCaseGlobals::check_upper_case(
                cx,
                "const parameter",
                Some(param.def_id),
                &param.name.ident(),
            );
        }
    }
}

#[cfg(test)]
mod tests;