clippy_lints/src/strings.rs

use clippy_utils::diagnostics::{span_lint, span_lint_and_sugg, span_lint_and_then};
use clippy_utils::res::{MaybeDef as _, MaybeQPath as _};
use clippy_utils::source::{snippet, snippet_with_applicability, snippet_with_context};
use clippy_utils::{SpanlessEq, get_expr_use_or_unification_node, get_parent_expr, is_lint_allowed, sym};
use rustc_attr_ir::lang_items::LangItem;
use rustc_errors::Applicability;
use rustc_hir::def::{DefKind, Res};
use rustc_hir::{BinOpKind, Expr, ExprKind, Node};
use rustc_lint::{LateContext, LateLintPass, LintContext as _, declare_lint_pass};
use rustc_middle::ty;

declare_clippy_lint! {
    /// ### What it does
    /// This lint checks for `.to_string()` method calls on values of type `&str`.
    ///
    /// ### Why restrict this?
    /// The `to_string` method is also used on other types to convert them to a string.
    /// When called on a `&str` it turns the `&str` into the owned variant `String`, which can be
    /// more specifically expressed with `.to_owned()`.
    ///
    /// ### Example
    /// ```no_run
    /// let _ = "str".to_string();
    /// ```
    /// Use instead:
    /// ```no_run
    /// let _ = "str".to_owned();
    /// ```
    #[clippy::version = "pre 1.29.0"]
    pub STR_TO_STRING,
    restriction,
    "using `to_string()` on a `&str`, which should be `to_owned()`"
}

declare_clippy_lint! {
    /// ### What it does
    /// Checks for all instances of `x + _` where `x` is of type
    /// `String`, but only if [`string_add_assign`](#string_add_assign) does *not*
    /// match.
    ///
    /// ### Why restrict this?
    /// This particular
    /// `Add` implementation is asymmetric (the other operand need not be `String`,
    /// but `x` does), while addition as mathematically defined is symmetric, and
    /// the `String::push_str(_)` function is a perfectly good replacement.
    /// Therefore, some dislike it and wish not to have it in their code.
    ///
    /// That said, other people think that string addition, having a long tradition
    /// in other languages is actually fine, which is why we decided to make this
    /// particular lint `allow` by default.
    ///
    /// ### Example
    /// ```no_run
    /// let x = "Hello".to_owned();
    /// x + ", World";
    /// ```
    ///
    /// Use instead:
    /// ```no_run
    /// let mut x = "Hello".to_owned();
    /// x.push_str(", World");
    /// ```
    #[clippy::version = "pre 1.29.0"]
    pub STRING_ADD,
    restriction,
    "using `x + ..` where x is a `String` instead of `push_str()`"
}

declare_clippy_lint! {
    /// ### What it does
    /// Checks for string appends of the form `x = x + y` (without
    /// `let`!).
    ///
    /// ### Why is this bad?
    /// It's not really bad, but some people think that the
    /// `.push_str(_)` method is more readable.
    ///
    /// ### Example
    /// ```no_run
    /// let mut x = "Hello".to_owned();
    /// x = x + ", World";
    ///
    /// // More readable
    /// x += ", World";
    /// x.push_str(", World");
    /// ```
    #[clippy::version = "pre 1.29.0"]
    pub STRING_ADD_ASSIGN,
    pedantic,
    "using `x = x + ..` where x is a `String` instead of `push_str()`"
}

declare_clippy_lint! {
    /// ### What it does
    /// Checks for the `as_bytes` method called on string literals
    /// that contain only ASCII characters.
    ///
    /// ### Why is this bad?
    /// Byte string literals (e.g., `b"foo"`) can be used
    /// instead. They are shorter but less discoverable than `as_bytes()`.
    ///
    /// ### Known problems
    /// `"str".as_bytes()` and the suggested replacement of `b"str"` are not
    /// equivalent because they have different types. The former is `&[u8]`
    /// while the latter is `&[u8; 3]`. That means in general they will have a
    /// different set of methods and different trait implementations.
    ///
    /// ```compile_fail
    /// fn f(v: Vec<u8>) {}
    ///
    /// f("...".as_bytes().to_owned()); // works
    /// f(b"...".to_owned()); // does not work, because arg is [u8; 3] not Vec<u8>
    ///
    /// fn g(r: impl std::io::Read) {}
    ///
    /// g("...".as_bytes()); // works
    /// g(b"..."); // does not work
    /// ```
    ///
    /// The actual equivalent of `"str".as_bytes()` with the same type is not
    /// `b"str"` but `&b"str"[..]`, which is a great deal of punctuation and not
    /// more readable than a function call.
    ///
    /// ### Example
    /// ```no_run
    /// let bstr = "a byte string".as_bytes();
    /// ```
    ///
    /// Use instead:
    /// ```no_run
    /// let bstr = b"a byte string";
    /// ```
    #[clippy::version = "pre 1.29.0"]
    pub STRING_LIT_AS_BYTES,
    nursery,
    "calling `as_bytes` on a string literal instead of using a byte string literal"
}

declare_clippy_lint! {
    /// ### What it does
    /// Checks for slice operations on strings
    ///
    /// ### Why restrict this?
    /// UTF-8 characters span multiple bytes, and it is easy to inadvertently confuse character
    /// counts and string indices. This may lead to panics, and should warrant some test cases
    /// containing wide UTF-8 characters. This lint is most useful in code that should avoid
    /// panics at all costs.
    ///
    /// ### Known problems
    /// Probably lots of false positives. If an index comes from a known valid position (e.g.
    /// obtained via `char_indices` over the same string), it is totally OK.
    ///
    /// ### Example
    /// ```rust,should_panic
    /// &"Ölkanne"[1..];
    /// ```
    #[clippy::version = "1.58.0"]
    pub STRING_SLICE,
    restriction,
    "slicing a string"
}

declare_lint_pass!(StrToString => [STR_TO_STRING]);

declare_lint_pass!(StringAdd => [STRING_ADD, STRING_ADD_ASSIGN, STRING_SLICE]);

declare_lint_pass!(StringLitAsBytes => [STRING_LIT_AS_BYTES]);

impl<'tcx> LateLintPass<'tcx> for StringAdd {
    fn check_expr(&mut self, cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) {
        match e.kind {
            ExprKind::Binary(op, lhs, _)
                if let BinOpKind::Add = op.node
                    && cx.typeck_results().expr_ty(lhs).is_lang_item(cx, LangItem::String)
                    && let ctxt = e.span.ctxt()
                    && op.span.ctxt() == ctxt
                    && !ctxt.in_external_macro(cx.tcx.sess.source_map()) =>
            {
                if !is_lint_allowed(cx, STRING_ADD_ASSIGN, e.hir_id)
                    && let Node::Expr(parent) = cx.tcx.parent_hir_node(e.hir_id)
                    && let ExprKind::Assign(assign_lhs, ..) = parent.kind
                    && parent.span.ctxt() == ctxt
                    && SpanlessEq::new(cx).eq_expr(ctxt, assign_lhs, lhs)
                {
                    return;
                }
                span_lint(
                    cx,
                    STRING_ADD,
                    e.span,
                    "you added something to a string. Consider using `String::push_str()` instead",
                );
            },
            ExprKind::Assign(lhs, rhs, _)
                if let ExprKind::Binary(op, add_lhs, _) = rhs.kind
                    && let BinOpKind::Add = op.node
                    && cx.typeck_results().expr_ty(lhs).is_lang_item(cx, LangItem::String)
                    && let ctxt = e.span.ctxt()
                    && SpanlessEq::new(cx).eq_expr(ctxt, lhs, add_lhs)
                    && rhs.span.ctxt() == ctxt
                    && op.span.ctxt() == ctxt
                    && !ctxt.in_external_macro(cx.tcx.sess.source_map()) =>
            {
                span_lint(
                    cx,
                    STRING_ADD_ASSIGN,
                    e.span,
                    "you assigned the result of adding something to this string. Consider using \
                         `String::push_str()` instead",
                );
            },
            ExprKind::Index(base, ..)
                if let ty::Ref(_, ty, _) = *cx.typeck_results().expr_ty_adjusted(base).kind()
                    && match *ty.kind() {
                        ty::Adt(def, _) => def.is_lang_item(cx, LangItem::String),
                        ty::Str => true,
                        _ => false,
                    }
                    && !e.span.in_external_macro(cx.tcx.sess.source_map()) =>
            {
                span_lint(
                    cx,
                    STRING_SLICE,
                    e.span,
                    "indexing into a string may panic if the index is within a UTF-8 character",
                );
            },
            _ => {},
        }
    }
}

// Max length a b"foo" string can take
const MAX_LENGTH_BYTE_STRING_LIT: usize = 32;

impl<'tcx> LateLintPass<'tcx> for StringLitAsBytes {
    fn check_expr(&mut self, cx: &LateContext<'tcx>, e: &'tcx Expr<'_>) {
        use rustc_ast::LitKind;

        if let ExprKind::MethodCall(path, receiver, ..) = &e.kind
            && path.ident.name == sym::as_bytes
            && let ExprKind::Lit(lit) = &receiver.kind
            && let LitKind::Str(lit_content, _) = &lit.node
            && !e.span.in_external_macro(cx.sess().source_map())
        {
            let callsite = snippet(cx, receiver.span.source_callsite(), r#""foo""#);
            let mut applicability = Applicability::MachineApplicable;
            if callsite.starts_with("include_str!") {
                span_lint_and_sugg(
                    cx,
                    STRING_LIT_AS_BYTES,
                    e.span,
                    "calling `as_bytes()` on `include_str!(..)`",
                    "consider using `include_bytes!(..)` instead",
                    snippet_with_applicability(cx, receiver.span.source_callsite(), r#""foo""#, &mut applicability)
                        .replacen("include_str", "include_bytes", 1),
                    applicability,
                );
            } else if lit_content.as_str().is_ascii()
                && lit_content.as_str().len() <= MAX_LENGTH_BYTE_STRING_LIT
                && !receiver.span.from_expansion()
            {
                if let Some((parent, id)) = get_expr_use_or_unification_node(cx.tcx, e)
                    && let Node::Expr(parent) = parent
                    && let ExprKind::Match(scrutinee, ..) = parent.kind
                    && scrutinee.hir_id == id
                {
                    // Don't lint. Byte strings produce `&[u8; N]` whereas `as_bytes()` produces
                    // `&[u8]`. This change would prevent matching with different sized slices.
                } else if !callsite.starts_with("env!") {
                    span_lint_and_sugg(
                        cx,
                        STRING_LIT_AS_BYTES,
                        e.span,
                        "calling `as_bytes()` on a string literal",
                        "consider using a byte string literal instead",
                        format!(
                            "b{}",
                            snippet_with_applicability(cx, receiver.span, r#""foo""#, &mut applicability)
                        ),
                        applicability,
                    );
                }
            }
        }

        if let ExprKind::MethodCall(path, recv, [], _) = &e.kind
            && path.ident.name == sym::into_bytes
            && let ExprKind::MethodCall(path, recv, [], _) = &recv.kind
            && matches!(path.ident.name, sym::to_owned | sym::to_string)
            && let ExprKind::Lit(lit) = &recv.kind
            && let LitKind::Str(lit_content, _) = &lit.node
            && lit_content.as_str().is_ascii()
            && lit_content.as_str().len() <= MAX_LENGTH_BYTE_STRING_LIT
            && !recv.span.from_expansion()
        {
            let mut applicability = Applicability::MachineApplicable;

            span_lint_and_sugg(
                cx,
                STRING_LIT_AS_BYTES,
                e.span,
                "calling `into_bytes()` on a string literal",
                "consider using a byte string literal instead",
                format!(
                    "b{}.to_vec()",
                    snippet_with_applicability(cx, recv.span, r#""..""#, &mut applicability)
                ),
                applicability,
            );
        }
    }
}

impl<'tcx> LateLintPass<'tcx> for StrToString {
    fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &Expr<'_>) {
        if expr.span.from_expansion() {
            return;
        }

        if let ExprKind::MethodCall(path, self_arg, [], _) = &expr.kind
            && path.ident.name == sym::to_string
            && let ty = cx.typeck_results().expr_ty(self_arg)
            && let ty::Ref(_, ty, ..) = ty.kind()
            && ty.is_str()
        {
            span_lint_and_then(
                cx,
                STR_TO_STRING,
                expr.span,
                "`to_string()` called on a `&str`",
                |diag| {
                    let mut applicability = Applicability::MachineApplicable;
                    let (snippet, _) =
                        snippet_with_context(cx, self_arg.span, expr.span.ctxt(), "..", &mut applicability);
                    diag.span_suggestion(expr.span, "try", format!("{snippet}.to_owned()"), applicability);
                },
            );
        } else if let ExprKind::Path(_) = expr.kind
            && let Some(parent) = get_parent_expr(cx, expr)
            && let ExprKind::Call(_, args) | ExprKind::MethodCall(_, _, args, _) = &parent.kind
            && args.iter().any(|a| a.hir_id == expr.hir_id)
            && let Res::Def(DefKind::AssocFn, def_id) = expr.res(cx)
            && cx.tcx.is_diagnostic_item(sym::to_string_method, def_id)
            && let Some(args) = cx.typeck_results().node_args_opt(expr.hir_id)
            && args.type_at(0).is_str()
        {
            // Detected `ToString::to_string` passed as an argument (generic: any call or method call)
            span_lint_and_sugg(
                cx,
                STR_TO_STRING,
                expr.span,
                "`ToString::to_string` used as `&str` to `String` converter",
                "try",
                "str::to_owned".to_string(),
                Applicability::MachineApplicable,
            );
        }
    }
}