diff --git a/src/librustdoc/html/render/print_item.rs b/src/librustdoc/html/render/print_item.rs index d2dc47af7ac4f..b559c427f6593 100644 --- a/src/librustdoc/html/render/print_item.rs +++ b/src/librustdoc/html/render/print_item.rs @@ -1474,34 +1474,27 @@ fn item_constant(w: &mut Buffer, cx: &mut Context<'_>, it: &clean::Item, c: &cle typ = c.type_.print(cx), ); - // FIXME: The code below now prints - // ` = _; // 100i32` - // if the expression is - // `50 + 50` - // which looks just wrong. - // Should we print - // ` = 100i32;` - // instead? - let value = c.value(tcx); let is_literal = c.is_literal(tcx); let expr = c.expr(tcx); - if value.is_some() || is_literal { + + if is_literal { write!(w, " = {expr};", expr = Escape(&expr)); + } else if let Some(ref value) = value { + write!(w, " = {value};", value = Escape(value)); } else { w.write_str(";"); } - if !is_literal { - if let Some(value) = &value { - let value_lowercase = value.to_lowercase(); - let expr_lowercase = expr.to_lowercase(); + let value_lowercase = value.as_ref().map(|s| s.to_lowercase()); + let expr_lowercase = Some(expr.to_lowercase()); - if value_lowercase != expr_lowercase - && value_lowercase.trim_end_matches("i32") != expr_lowercase - { - write!(w, " // {value}", value = Escape(value)); - } + if value_lowercase != expr_lowercase + && value_lowercase.as_ref().map(|s| s.trim_end_matches("i32")) + != expr_lowercase.as_deref() + { + if let Some(ref value) = value { + write!(w, " // {value}", value = Escape(value.as_str())); } } }); diff --git a/src/tools/clippy/clippy_lints/src/invalid_utf8_in_unchecked.rs b/src/tools/clippy/clippy_lints/src/invalid_utf8_in_unchecked.rs new file mode 100644 index 0000000000000..6a4861747d267 --- /dev/null +++ b/src/tools/clippy/clippy_lints/src/invalid_utf8_in_unchecked.rs @@ -0,0 +1,74 @@ +use clippy_utils::diagnostics::span_lint; +use clippy_utils::{match_function_call, paths}; +use rustc_ast::{BorrowKind, LitKind}; +use rustc_hir::{Expr, ExprKind}; +use rustc_lint::{LateContext, LateLintPass}; +use rustc_session::{declare_lint_pass, declare_tool_lint}; +use rustc_span::source_map::Spanned; +use rustc_span::Span; + +declare_clippy_lint! { + /// ### What it does + /// Checks for `std::str::from_utf8_unchecked` with an invalid UTF-8 literal + /// + /// ### Why is this bad? + /// Creating such a `str` would result in undefined behavior + /// + /// ### Example + /// ```rust + /// # #[allow(unused)] + /// unsafe { + /// std::str::from_utf8_unchecked(b"cl\x82ippy"); + /// } + /// ``` + #[clippy::version = "1.64.0"] + pub INVALID_UTF8_IN_UNCHECKED, + correctness, + "using a non UTF-8 literal in `std::std::from_utf8_unchecked`" +} +declare_lint_pass!(InvalidUtf8InUnchecked => [INVALID_UTF8_IN_UNCHECKED]); + +impl<'tcx> LateLintPass<'tcx> for InvalidUtf8InUnchecked { + fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'tcx>) { + if let Some([arg]) = match_function_call(cx, expr, &paths::STR_FROM_UTF8_UNCHECKED) { + match &arg.kind { + ExprKind::Lit(Spanned { node: lit, .. }) => { + if let LitKind::ByteStr(bytes, _) = &lit + && std::str::from_utf8(bytes).is_err() + { + lint(cx, expr.span); + } + }, + ExprKind::AddrOf(BorrowKind::Ref, _, Expr { kind: ExprKind::Array(args), .. }) => { + let elements = args.iter().map(|e|{ + match &e.kind { + ExprKind::Lit(Spanned { node: lit, .. }) => match lit { + LitKind::Byte(b) => Some(*b), + #[allow(clippy::cast_possible_truncation)] + LitKind::Int(b, _) => Some(*b as u8), + _ => None + } + _ => None + } + }).collect::>>(); + + if let Some(elements) = elements + && std::str::from_utf8(&elements).is_err() + { + lint(cx, expr.span); + } + } + _ => {} + } + } + } +} + +fn lint(cx: &LateContext<'_>, span: Span) { + span_lint( + cx, + INVALID_UTF8_IN_UNCHECKED, + span, + "non UTF-8 literal in `std::str::from_utf8_unchecked`", + ); +} diff --git a/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.rs b/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.rs new file mode 100644 index 0000000000000..3dc096d3197fb --- /dev/null +++ b/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.rs @@ -0,0 +1,20 @@ +#![warn(clippy::invalid_utf8_in_unchecked)] + +fn main() { + // Valid + unsafe { + std::str::from_utf8_unchecked(&[99, 108, 105, 112, 112, 121]); + std::str::from_utf8_unchecked(&[b'c', b'l', b'i', b'p', b'p', b'y']); + std::str::from_utf8_unchecked(b"clippy"); + + let x = 0xA0; + std::str::from_utf8_unchecked(&[0xC0, x]); + } + + // Invalid + unsafe { + std::str::from_utf8_unchecked(&[99, 108, 130, 105, 112, 112, 121]); + std::str::from_utf8_unchecked(&[b'c', b'l', b'\x82', b'i', b'p', b'p', b'y']); + std::str::from_utf8_unchecked(b"cl\x82ippy"); + } +} diff --git a/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.stderr b/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.stderr new file mode 100644 index 0000000000000..c89cd2758ee9f --- /dev/null +++ b/src/tools/clippy/tests/ui/invalid_utf8_in_unchecked.stderr @@ -0,0 +1,22 @@ +error: non UTF-8 literal in `std::str::from_utf8_unchecked` + --> $DIR/invalid_utf8_in_unchecked.rs:16:9 + | +LL | std::str::from_utf8_unchecked(&[99, 108, 130, 105, 112, 112, 121]); + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + | + = note: `-D clippy::invalid-utf8-in-unchecked` implied by `-D warnings` + +error: non UTF-8 literal in `std::str::from_utf8_unchecked` + --> $DIR/invalid_utf8_in_unchecked.rs:17:9 + | +LL | std::str::from_utf8_unchecked(&[b'c', b'l', b'/x82', b'i', b'p', b'p', b'y']); + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +error: non UTF-8 literal in `std::str::from_utf8_unchecked` + --> $DIR/invalid_utf8_in_unchecked.rs:18:9 + | +LL | std::str::from_utf8_unchecked(b"cl/x82ippy"); + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +error: aborting due to 3 previous errors +