A new lint for shared code in if blocks

* Added expression check for shared_code_in_if_blocks
* Finishing touches for the shared_code_in_if_blocks lint
* Applying PR suggestions
* Update lints yay
* Moved test into subfolder

Author: xFrednet

CHANGELOG.md

@@ -2455,6 +2455,7 @@ Released 2018-09-13
 [`shadow_reuse`]: https://rust-lang.github.io/rust-clippy/master/index.html#shadow_reuse
 [`shadow_same`]: https://rust-lang.github.io/rust-clippy/master/index.html#shadow_same
 [`shadow_unrelated`]: https://rust-lang.github.io/rust-clippy/master/index.html#shadow_unrelated
+[`shared_code_in_if_blocks`]: https://rust-lang.github.io/rust-clippy/master/index.html#shared_code_in_if_blocks
 [`short_circuit_statement`]: https://rust-lang.github.io/rust-clippy/master/index.html#short_circuit_statement
 [`should_assert_eq`]: https://rust-lang.github.io/rust-clippy/master/index.html#should_assert_eq
 [`should_implement_trait`]: https://rust-lang.github.io/rust-clippy/master/index.html#should_implement_trait

clippy_lints/src/copies.rs

@@ -3,7 +3,10 @@ use clippy_utils::{eq_expr_value, in_macro, search_same, SpanlessEq, SpanlessHas
 use clippy_utils::{get_parent_expr, if_sequence};
 use rustc_hir::{Block, Expr, ExprKind};
 use rustc_lint::{LateContext, LateLintPass};
+use rustc_middle::hir::map::Map;
 use rustc_session::{declare_lint_pass, declare_tool_lint};
+use rustc_span::source_map::Span;
+use std::borrow::Cow;
 
 declare_clippy_lint! {
     /// **What it does:** Checks for consecutive `if`s with the same condition.
@@ -104,7 +107,45 @@ declare_clippy_lint! {
     "`if` with the same `then` and `else` blocks"
 }
 
-declare_lint_pass!(CopyAndPaste => [IFS_SAME_COND, SAME_FUNCTIONS_IN_IF_CONDITION, IF_SAME_THEN_ELSE]);
+declare_clippy_lint! {
+    /// **What it does:** Checks if the `if` and `else` block contain shared code that can be
+    /// moved out of the blocks.
+    ///
+    /// **Why is this bad?** Duplicate code is less maintainable.
+    ///
+    /// **Known problems:** Hopefully none.
+    ///
+    /// **Example:**
+    /// ```ignore
+    /// let foo = if … {
+    ///     println!("Hello World");
+    ///     13
+    /// } else {
+    ///     println!("Hello World");
+    ///     42
+    /// };
+    /// ```
+    ///
+    /// Could be written as:
+    /// ```ignore
+    /// println!("Hello World");
+    /// let foo = if … {
+    ///     13
+    /// } else {
+    ///     42
+    /// };
+    /// ```
+    pub SHARED_CODE_IN_IF_BLOCKS,
+    pedantic,
+    "`if` statement with shared code in all blocks"
+}
+
+declare_lint_pass!(CopyAndPaste => [
+    IFS_SAME_COND,
+    SAME_FUNCTIONS_IN_IF_CONDITION,
+    IF_SAME_THEN_ELSE,
+    SHARED_CODE_IN_IF_BLOCKS
+]);
 
 impl<'tcx> LateLintPass<'tcx> for CopyAndPaste {
     fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>) {
@@ -121,30 +162,256 @@ impl<'tcx> LateLintPass<'tcx> for CopyAndPaste {
             }
 
             let (conds, blocks) = if_sequence(expr);
-            lint_same_then_else(cx, &blocks);
+            // Conditions
             lint_same_cond(cx, &conds);
             lint_same_fns_in_if_cond(cx, &conds);
+            // Block duplication
+            lint_same_then_else(cx, &blocks, conds.len() != blocks.len(), expr);
         }
     }
 }
 
-/// Implementation of `IF_SAME_THEN_ELSE`.
-fn lint_same_then_else(cx: &LateContext<'_>, blocks: &[&Block<'_>]) {
-    let eq: &dyn Fn(&&Block<'_>, &&Block<'_>) -> bool =
-        &|&lhs, &rhs| -> bool { SpanlessEq::new(cx).eq_block(lhs, rhs) };
+/// Implementation of `SHARED_CODE_IN_IF_BLOCKS` and `IF_SAME_THEN_ELSE` if the blocks are equal.
+fn lint_same_then_else<'tcx>(
+    cx: &LateContext<'tcx>,
+    blocks: &[&Block<'tcx>],
+    has_unconditional_else: bool,
+    expr: &'tcx Expr<'_>,
+) {
+    // We only lint ifs with multiple blocks
+    // TODO xFrednet 2021-01-01: Check if it's an else if block
+    if blocks.len() < 2 {
+        return;
+    }
 
-    if let Some((i, j)) = search_same_sequenced(blocks, eq) {
-        span_lint_and_note(
+    let has_expr = blocks[0].expr.is_some();
+
+    // Check if each block has shared code
+    let mut start_eq = usize::MAX;
+    let mut end_eq = usize::MAX;
+    let mut expr_eq = true;
+    for (index, win) in blocks.windows(2).enumerate() {
+        let l_stmts = win[0].stmts;
+        let r_stmts = win[1].stmts;
+
+        let mut evaluator = SpanlessEq::new(cx);
+        let current_start_eq = count_eq(&mut l_stmts.iter(), &mut r_stmts.iter(), |l, r| evaluator.eq_stmt(l, r));
+        let current_end_eq = count_eq(&mut l_stmts.iter().rev(), &mut r_stmts.iter().rev(), |l, r| {
+            evaluator.eq_stmt(l, r)
+        });
+        let block_expr_eq = both(&win[0].expr, &win[1].expr, |l, r| evaluator.eq_expr(l, r));
+
+        // IF_SAME_THEN_ELSE
+        // We only lint the first two blocks (index == 0). Further blocks will be linted when that if
+        // statement is checked
+        if index == 0 && block_expr_eq && l_stmts.len() == r_stmts.len() && l_stmts.len() == current_start_eq {
+            span_lint_and_note(
+                cx,
+                IF_SAME_THEN_ELSE,
+                win[0].span,
+                "this `if` has identical blocks",
+                Some(win[1].span),
+                "same as this",
+            );
+
+            return;
+        }
+
+        start_eq = start_eq.min(current_start_eq);
+        end_eq = end_eq.min(current_end_eq);
+        expr_eq &= block_expr_eq;
+
+        // We can return if the eq count is 0 from both sides or if it has no unconditional else case
+        if !has_unconditional_else || (start_eq == 0 && end_eq == 0 && (has_expr && !expr_eq)) {
+            return;
+        }
+    }
+
+    if has_expr && !expr_eq {
+        end_eq = 0;
+    }
+
+    // Check if the regions are overlapping. Set `end_eq` to prevent the overlap
+    let min_block_size = blocks.iter().map(|x| x.stmts.len()).min().unwrap();
+    if (start_eq + end_eq) > min_block_size {
+        end_eq = min_block_size - start_eq;
+    }
+
+    // Only the start is the same
+    if start_eq != 0 && end_eq == 0 && (!has_expr || !expr_eq) {
+        emit_shared_code_in_if_blocks_lint(cx, start_eq, 0, false, blocks, expr);
+    } else if end_eq != 0 && (!has_expr || !expr_eq) {
+        let block = blocks[blocks.len() - 1];
+        let stmts = block.stmts.split_at(start_eq).1;
+        let (block_stmts, moved_stmts) = stmts.split_at(stmts.len() - end_eq);
+
+        // Scan block
+        let mut walker = SymbolFinderVisitor::new(cx);
+        for stmt in block_stmts {
+            intravisit::walk_stmt(&mut walker, stmt);
+        }
+        let mut block_defs = walker.defs;
+
+        // Scan moved stmts
+        let mut moved_start: Option<usize> = None;
+        let mut walker = SymbolFinderVisitor::new(cx);
+        for (index, stmt) in moved_stmts.iter().enumerate() {
+            intravisit::walk_stmt(&mut walker, stmt);
+
+            for value in &walker.uses {
+                // Well we can't move this and all prev statements. So reset
+                if block_defs.contains(&value) {
+                    moved_start = Some(index + 1);
+                    walker.defs.drain().for_each(|x| {
+                        block_defs.insert(x);
+                    });
+                }
+            }
+
+            walker.uses.clear();
+        }
+
+        if let Some(moved_start) = moved_start {
+            end_eq -= moved_start;
+        }
+
+        let mut end_linable = true;
+        if let Some(expr) = block.expr {
+            intravisit::walk_expr(&mut walker, expr);
+            end_linable = walker.uses.iter().any(|x| !block_defs.contains(x));
+        }
+
+        emit_shared_code_in_if_blocks_lint(cx, start_eq, end_eq, end_linable, blocks, expr);
+    }
+}
+
+fn emit_shared_code_in_if_blocks_lint(
+    cx: &LateContext<'tcx>,
+    start_stmts: usize,
+    end_stmts: usize,
+    lint_end: bool,
+    blocks: &[&Block<'tcx>],
+    if_expr: &'tcx Expr<'_>,
+) {
+    if start_stmts == 0 && !lint_end {
+        return;
+    }
+
+    // (help, span, suggestion)
+    let mut suggestions: Vec<(&str, Span, String)> = vec![];
+
+    if start_stmts > 0 {
+        let block = blocks[0];
+        let span_start = first_line_of_span(cx, if_expr.span).shrink_to_lo();
+        let span_end = block.stmts[start_stmts - 1].span.source_callsite();
+
+        let cond_span = first_line_of_span(cx, if_expr.span).until(block.span);
+        let cond_snippet = reindent_multiline(snippet(cx, cond_span, "_"), false, None);
+        let cond_indent = indent_of(cx, cond_span);
+        let moved_span = block.stmts[0].span.source_callsite().to(span_end);
+        let moved_snippet = reindent_multiline(snippet(cx, moved_span, "_"), true, None);
+        let suggestion = moved_snippet.to_string() + "\n" + &cond_snippet + "{";
+        let suggestion = reindent_multiline(Cow::Borrowed(&suggestion), true, cond_indent);
+
+        let span = span_start.to(span_end);
+        suggestions.push(("START HELP", span, suggestion.to_string()));
+    }
+
+    if lint_end {
+        let block = blocks[blocks.len() - 1];
+        let span_end = block.span.shrink_to_hi();
+
+        let moved_start = if end_stmts == 0 && block.expr.is_some() {
+            block.expr.unwrap().span
+        } else {
+            block.stmts[block.stmts.len() - end_stmts].span
+        }
+        .source_callsite();
+        let moved_end = if let Some(expr) = block.expr {
+            expr.span
+        } else {
+            block.stmts[block.stmts.len() - 1].span
+        }
+        .source_callsite();
+
+        let moved_span = moved_start.to(moved_end);
+        let moved_snipped = reindent_multiline(snippet(cx, moved_span, "_"), true, None);
+        let indent = indent_of(cx, if_expr.span.shrink_to_hi());
+        let suggestion = "}\n".to_string() + &moved_snipped;
+        let suggestion = reindent_multiline(Cow::Borrowed(&suggestion), true, indent);
+
+        let span = moved_start.to(span_end);
+        suggestions.push(("END_RANGE", span, suggestion.to_string()));
+    }
+
+    if suggestions.len() == 1 {
+        let (_, span, sugg) = &suggestions[0];
+        span_lint_and_sugg(
             cx,
-            IF_SAME_THEN_ELSE,
-            j.span,
-            "this `if` has identical blocks",
-            Some(i.span),
-            "same as this",
+            SHARED_CODE_IN_IF_BLOCKS,
+            *span,
+            "All code blocks contain the same code",
+            "Consider moving the code out like this",
+            sugg.clone(),
+            Applicability::Unspecified,
+        );
+    } else {
+        span_lint_and_then(
+            cx,
+            SHARED_CODE_IN_IF_BLOCKS,
+            if_expr.span,
+            "All if blocks contain the same code",
+            move |diag| {
+                for (help, span, sugg) in suggestions {
+                    diag.span_suggestion(span, help, sugg, Applicability::Unspecified);
+                }
+            },
         );
     }
 }
 
+pub struct SymbolFinderVisitor<'a, 'tcx> {
+    cx: &'a LateContext<'tcx>,
+    defs: FxHashSet<HirId>,
+    uses: FxHashSet<HirId>,
+}
+
+impl<'a, 'tcx> SymbolFinderVisitor<'a, 'tcx> {
+    fn new(cx: &'a LateContext<'tcx>) -> Self {
+        SymbolFinderVisitor {
+            cx,
+            defs: FxHashSet::default(),
+            uses: FxHashSet::default(),
+        }
+    }
+}
+
+impl<'a, 'tcx> Visitor<'tcx> for SymbolFinderVisitor<'a, 'tcx> {
+    type Map = Map<'tcx>;
+
+    fn nested_visit_map(&mut self) -> NestedVisitorMap<Self::Map> {
+        NestedVisitorMap::All(self.cx.tcx.hir())
+    }
+
+    fn visit_local(&mut self, l: &'tcx rustc_hir::Local<'tcx>) {
+        let local_id = l.pat.hir_id;
+        self.defs.insert(local_id);
+        if let Some(expr) = l.init {
+            intravisit::walk_expr(self, expr);
+        }
+    }
+
+    fn visit_qpath(&mut self, qpath: &'tcx rustc_hir::QPath<'tcx>, id: HirId, _span: rustc_span::Span) {
+        if let rustc_hir::QPath::Resolved(_, ref path) = *qpath {
+            if path.segments.len() == 1 {
+                if let rustc_hir::def::Res::Local(var) = self.cx.qpath_res(qpath, id) {
+                    self.uses.insert(var);
+                }
+            }
+        }
+    }
+}
+
 /// Implementation of `IFS_SAME_COND`.
 fn lint_same_cond(cx: &LateContext<'_>, conds: &[&Expr<'_>]) {
     let hash: &dyn Fn(&&Expr<'_>) -> u64 = &|expr| -> u64 {
@@ -198,15 +465,3 @@ fn lint_same_fns_in_if_cond(cx: &LateContext<'_>, conds: &[&Expr<'_>]) {
         );
     }
 }
-
-fn search_same_sequenced<T, Eq>(exprs: &[T], eq: Eq) -> Option<(&T, &T)>
-where
-    Eq: Fn(&T, &T) -> bool,
-{
-    for win in exprs.windows(2) {
-        if eq(&win[0], &win[1]) {
-            return Some((&win[0], &win[1]));
-        }
-    }
-    None
-}

clippy_lints/src/lib.rs

@@ -617,6 +617,7 @@ pub fn register_plugins(store: &mut rustc_lint::LintStore, sess: &Session, conf:
         &copies::IFS_SAME_COND,
         &copies::IF_SAME_THEN_ELSE,
         &copies::SAME_FUNCTIONS_IN_IF_CONDITION,
+        &copies::SHARED_CODE_IN_IF_BLOCKS,
         &copy_iterator::COPY_ITERATOR,
         &create_dir::CREATE_DIR,
         &dbg_macro::DBG_MACRO,
@@ -1366,6 +1367,7 @@ pub fn register_plugins(store: &mut rustc_lint::LintStore, sess: &Session, conf:
         LintId::of(&casts::PTR_AS_PTR),
         LintId::of(&checked_conversions::CHECKED_CONVERSIONS),
         LintId::of(&copies::SAME_FUNCTIONS_IN_IF_CONDITION),
+        LintId::of(&copies::SHARED_CODE_IN_IF_BLOCKS),
         LintId::of(&copy_iterator::COPY_ITERATOR),
         LintId::of(&default::DEFAULT_TRAIT_ACCESS),
         LintId::of(&dereference::EXPLICIT_DEREF_METHODS),