[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-pointer-arithmetic #127394

flovent · 2025-02-16T13:17:43Z

this PR fixes #126424
for ArraySubScriptExpr, hasBase Matcher will get right operand when it is not integer type, but is not for sure that left operand is integer type. For the example code below hasBase will get r for the Subsequent matching and causing false positive.

template <typename R>
int f(std::map<R*, int>& map, R* r) {
  return map[r];
}

so is needed to see if index is integer type to avoid this situation.

…ter-arithmetic

llvmbot · 2025-02-16T13:18:15Z

@llvm/pr-subscribers-clang-tidy

Author: None (flovent)

Changes

this PR fixs #126424
for ArraySubScriptExpr, hasBase Matcher will get right operand when it is not Integer type, but is not for sure that left operand is interger type. For the example code below hasBase will get r for the Subsequent matching and causing false positive.

template &lt;typename R&gt;
int f(std::map&lt;R*, int&gt;&amp; map, R* r) {
  return map[r];
}

so is needed to see if index is interger type to avoid this situation.

Full diff: https://github.com/llvm/llvm-project/pull/127394.diff

3 Files Affected:

(modified) clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp (+2-1)
(modified) clang-tools-extra/docs/ReleaseNotes.rst (+4)
(added) clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp (+19)

diff --git a/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp b/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
index a1494a095f5b6..0d68790349fb5 100644
--- a/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
+++ b/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
@@ -42,7 +42,8 @@ void ProBoundsPointerArithmeticCheck::registerMatchers(MatchFinder *Finder) {
       arraySubscriptExpr(
           hasBase(ignoringImpCasts(
               anyOf(AllPointerTypes,
-                    hasType(decayedType(hasDecayedType(pointerType())))))))
+                    hasType(decayedType(hasDecayedType(pointerType())))))),
+          hasIndex(hasType(isInteger())))
           .bind("expr"),
       this);
 }
diff --git a/clang-tools-extra/docs/ReleaseNotes.rst b/clang-tools-extra/docs/ReleaseNotes.rst
index 6b8fe22242417..937d710fe6100 100644
--- a/clang-tools-extra/docs/ReleaseNotes.rst
+++ b/clang-tools-extra/docs/ReleaseNotes.rst
@@ -110,6 +110,10 @@ Changes in existing checks
   <clang-tidy/checks/misc/redundant-expression>` check by providing additional
   examples and fixing some macro related false positives.
 
+- Improved :doc:`cppcoreguidelines-pro-bounds-pointer-arithmetic
+  <clang-tidy/checks/cppcoreguidelines/pro-bounds-pointer-arithmetic>` check by
+  fix false positives related to operator overloading and templates.
+
 Removed checks
 ^^^^^^^^^^^^^^
 
diff --git a/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp b/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp
new file mode 100644
index 0000000000000..b6b7a9664f38d
--- /dev/null
+++ b/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp
@@ -0,0 +1,19 @@
+// RUN: %check_clang_tidy %s cppcoreguidelines-pro-bounds-pointer-arithmetic %t
+
+namespace std {
+template <typename, typename>
+class pair {};
+
+template <typename Key, typename Value>
+class map {
+  public:
+   using value_type = pair<Key, Value>;
+   value_type& operator[](const Key& key);
+   value_type& operator[](Key&& key);
+ };
+}
+
+template <typename R>
+int f(std::map<R*, int>& map, R* r) {
+  return map[r]; // OK
+}
\ No newline at end of file

llvmbot · 2025-02-16T13:18:15Z

@llvm/pr-subscribers-clang-tools-extra

Author: None (flovent)

Changes

this PR fixs #126424
for ArraySubScriptExpr, hasBase Matcher will get right operand when it is not Integer type, but is not for sure that left operand is interger type. For the example code below hasBase will get r for the Subsequent matching and causing false positive.

template &lt;typename R&gt;
int f(std::map&lt;R*, int&gt;&amp; map, R* r) {
  return map[r];
}

so is needed to see if index is interger type to avoid this situation.

Full diff: https://github.com/llvm/llvm-project/pull/127394.diff

3 Files Affected:

(modified) clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp (+2-1)
(modified) clang-tools-extra/docs/ReleaseNotes.rst (+4)
(added) clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp (+19)

diff --git a/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp b/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
index a1494a095f5b6..0d68790349fb5 100644
--- a/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
+++ b/clang-tools-extra/clang-tidy/cppcoreguidelines/ProBoundsPointerArithmeticCheck.cpp
@@ -42,7 +42,8 @@ void ProBoundsPointerArithmeticCheck::registerMatchers(MatchFinder *Finder) {
       arraySubscriptExpr(
           hasBase(ignoringImpCasts(
               anyOf(AllPointerTypes,
-                    hasType(decayedType(hasDecayedType(pointerType())))))))
+                    hasType(decayedType(hasDecayedType(pointerType())))))),
+          hasIndex(hasType(isInteger())))
           .bind("expr"),
       this);
 }
diff --git a/clang-tools-extra/docs/ReleaseNotes.rst b/clang-tools-extra/docs/ReleaseNotes.rst
index 6b8fe22242417..937d710fe6100 100644
--- a/clang-tools-extra/docs/ReleaseNotes.rst
+++ b/clang-tools-extra/docs/ReleaseNotes.rst
@@ -110,6 +110,10 @@ Changes in existing checks
   <clang-tidy/checks/misc/redundant-expression>` check by providing additional
   examples and fixing some macro related false positives.
 
+- Improved :doc:`cppcoreguidelines-pro-bounds-pointer-arithmetic
+  <clang-tidy/checks/cppcoreguidelines/pro-bounds-pointer-arithmetic>` check by
+  fix false positives related to operator overloading and templates.
+
 Removed checks
 ^^^^^^^^^^^^^^
 
diff --git a/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp b/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp
new file mode 100644
index 0000000000000..b6b7a9664f38d
--- /dev/null
+++ b/clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-bounds-pointer-arithmetic-issue126424.cpp
@@ -0,0 +1,19 @@
+// RUN: %check_clang_tidy %s cppcoreguidelines-pro-bounds-pointer-arithmetic %t
+
+namespace std {
+template <typename, typename>
+class pair {};
+
+template <typename Key, typename Value>
+class map {
+  public:
+   using value_type = pair<Key, Value>;
+   value_type& operator[](const Key& key);
+   value_type& operator[](Key&& key);
+ };
+}
+
+template <typename R>
+int f(std::map<R*, int>& map, R* r) {
+  return map[r]; // OK
+}
\ No newline at end of file

clang-tools-extra/docs/ReleaseNotes.rst

flovent · 2025-02-24T03:28:47Z

ping

carlosgalvezp · 2025-04-30T19:03:46Z

LGTM in general, but won't this fail with a std::map<int, int>, since the key is an integer?

Please rebase on top of latest main and resolve conflicts.

carlosgalvezp · 2025-04-30T19:05:27Z

clang-tools-extra/docs/ReleaseNotes.rst

@@ -106,6 +106,10 @@ Changes in existing checks
  <clang-tidy/checks/bugprone/unsafe-functions>` check to allow specifying
  additional C++ member functions to match.

+- Improved :doc:`cppcoreguidelines-pro-bounds-pointer-arithmetic
+  <clang-tidy/checks/cppcoreguidelines/pro-bounds-pointer-arithmetic>` check by
+  fix false positives related to operator overloading and templates.


"by fixing"

Also, "operator overloading and templates" seems too generic, maybe "when calling indexing operators that do not perform pointer arithmetic, for example std::map::operator[]"?

I think template should be mentioned because it's the necessary condition in this false positive, for std::map::operator[] or other indexing operators like std::vector::operator[] outside of template, there will be CallExpr generated, this checker didn't match CallExpr before this PR's change.

Using indexing operator does make it more specific, since ArraySubScriptExpr only shows for this kind of overload operator.

Maybe "by fixing false positives when calling indexing operators that do not perform pointer arithmetic in template, for example std::map::operator[]"?

flovent · 2025-05-01T08:49:17Z

but won't this fail with a std::map<int, int>, since the key is an integer?

It won't, ArraySubScriptExpr is generated here because variable map's type is dependent, it's AST:

ArraySubscriptExpr <col:10, col:15> '<dependent type>' lvalue
      -DeclRefExpr <col:10> 'std::map<R *, int>':'map<R *, int>' lvalue ParmVar 0x227b3b48 'map' 'std::map<R *, int> &'
      -DeclRefExpr <col:14> 'R *' lvalue ParmVar 0x227b3bc8 'r' 'R *'

std::map's key is unrelated here, for std::map<int, int> a CallExpr will be generated since it's known that std::map<int, int> has a overload operator []

[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-poin…

1bd661b

…ter-arithmetic

llvmbot added clang-tools-extra clang-tidy labels Feb 16, 2025

add missing new line for testcase

ab8a217

EugeneZelenko requested review from PiotrZSL, 5chmidti, carlosgalvezp and HerrCai0907 February 16, 2025 15:49

EugeneZelenko reviewed Feb 16, 2025

View reviewed changes

clang-tools-extra/docs/ReleaseNotes.rst Show resolved Hide resolved

keep alpha order for release note

e898caf

carlosgalvezp reviewed Apr 30, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-pointer-arithmetic #127394

[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-pointer-arithmetic #127394

flovent commented Feb 16, 2025 •

edited by carlosgalvezp

Loading

llvmbot commented Feb 16, 2025

llvmbot commented Feb 16, 2025

flovent commented Feb 24, 2025

carlosgalvezp commented Apr 30, 2025

carlosgalvezp Apr 30, 2025 •

edited

Loading

flovent May 1, 2025 •

edited

Loading

flovent commented May 1, 2025

[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-pointer-arithmetic #127394

Are you sure you want to change the base?

[clang-tidy] Fix false positive for cppcoreguidelines-pro-bounds-pointer-arithmetic #127394

Conversation

flovent commented Feb 16, 2025 • edited by carlosgalvezp Loading

llvmbot commented Feb 16, 2025

llvmbot commented Feb 16, 2025

flovent commented Feb 24, 2025

carlosgalvezp commented Apr 30, 2025

carlosgalvezp Apr 30, 2025 • edited Loading

Choose a reason for hiding this comment

flovent May 1, 2025 • edited Loading

Choose a reason for hiding this comment

flovent commented May 1, 2025

flovent commented Feb 16, 2025 •

edited by carlosgalvezp

Loading

carlosgalvezp Apr 30, 2025 •

edited

Loading

flovent May 1, 2025 •

edited

Loading