[clang][test] Fix test issue under LLVM_REVERSE_ITERATION

[ilovepi]

The order of these items may not be the same if reverse iteration is
enabled. From what I can tell this is related to visitation order, and I
don't see an easy way to handle that using the typical solutions, like
MapVector, etc. For now, just use CHECK-DAG to get the test into a
passing state.

Fixes #167057

[ilovepi]

• [clang][test] Fix test issue under LLVM_REVERSE_ITERATION #167394 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-clang

Author: Paul Kirth (ilovepi)

Changes

The order of these items may not be the same if reverse iteration is
enabled. From what I can tell this is related to visitation order, and I
don't see an easy way to handle that using the typical solutions, like
MapVector, etc. For now, just use CHECK-DAG to get the test into a
passing state.

Fixes #167057

---

Full diff: https://github.com/llvm/llvm-project/pull/167394.diff

1 Files Affected:

• (modified) clang/test/Analysis/analyzeOneFunction.cpp (+3-3)

diff --git a/clang/test/Analysis/analyzeOneFunction.cpp b/clang/test/Analysis/analyzeOneFunction.cpp
index 3a362dfd9a08c..9b752e29320a4 100644
--- a/clang/test/Analysis/analyzeOneFunction.cpp
+++ b/clang/test/Analysis/analyzeOneFunction.cpp
@@ -5,9 +5,9 @@
 // RUN:   -analyze-function="c:@S@Window@F@overloaded#I#"
 
 // RUN: %clang_extdef_map %s | FileCheck %s
-// CHECK:      27:c:@S@Window@F@overloaded#I#
-// CHECK-NEXT: 27:c:@S@Window@F@overloaded#C#
-// CHECK-NEXT: 27:c:@S@Window@F@overloaded#d#
+// CHECK-DAG:      27:c:@S@Window@F@overloaded#I#
+// CHECK-DAG: 27:c:@S@Window@F@overloaded#C#
+// CHECK-DAG: 27:c:@S@Window@F@overloaded#d#
 
 void clang_analyzer_warnIfReached();
 

[llvmbot]

@llvm/pr-subscribers-clang-static-analyzer-1

Author: Paul Kirth (ilovepi)

Changes

The order of these items may not be the same if reverse iteration is
enabled. From what I can tell this is related to visitation order, and I
don't see an easy way to handle that using the typical solutions, like
MapVector, etc. For now, just use CHECK-DAG to get the test into a
passing state.

Fixes #167057

---

Full diff: https://github.com/llvm/llvm-project/pull/167394.diff

1 Files Affected:

• (modified) clang/test/Analysis/analyzeOneFunction.cpp (+3-3)

diff --git a/clang/test/Analysis/analyzeOneFunction.cpp b/clang/test/Analysis/analyzeOneFunction.cpp
index 3a362dfd9a08c..9b752e29320a4 100644
--- a/clang/test/Analysis/analyzeOneFunction.cpp
+++ b/clang/test/Analysis/analyzeOneFunction.cpp
@@ -5,9 +5,9 @@
 // RUN:   -analyze-function="c:@S@Window@F@overloaded#I#"
 
 // RUN: %clang_extdef_map %s | FileCheck %s
-// CHECK:      27:c:@S@Window@F@overloaded#I#
-// CHECK-NEXT: 27:c:@S@Window@F@overloaded#C#
-// CHECK-NEXT: 27:c:@S@Window@F@overloaded#d#
+// CHECK-DAG:      27:c:@S@Window@F@overloaded#I#
+// CHECK-DAG: 27:c:@S@Window@F@overloaded#C#
+// CHECK-DAG: 27:c:@S@Window@F@overloaded#d#
 
 void clang_analyzer_warnIfReached();
 

[necto]

The fix makes sense.
The order in which functions are analyzed does not matter for this test.

P.S. I was not aware of LLVM_REVERSE_ITERATION, seems like a great tool!

[necto]

FYI, CSA analyzes entry points in a reverse topological order of the call graph. There is a minimal, but nonzero number of dependencies between entry points. One notable example is the inline limit. If a callee was inlined when analyzing entry_point_a() it might no longer be inlined when analyzing subsequent entry_point_b(). So theoretically some other tests that are sensitive to a callee being inlined could start failing with LLVM_REVERSE_ITERATION=ON, but apparently that didn't happen so far.