Fix #14995 (Suppression file name cache) (#8684)

Implement a file name cache per suppression to avoid calling
PathMatch::match again and again for the same suppression and file name
combination.

Local benchmark running cppcheck with misra c-2012 has been performed.

**Before** 
10min 34sec
**After**
1min 36sec

In this test it was a few header files producing the majority of the
error messages, which is why the cache speeds it up by a lot.

Author: MartinBP

lib/suppressions.cpp

@@ -423,7 +423,7 @@ SuppressionList::Suppression::Result SuppressionList::Suppression::isSuppressed(
             if (!thisAndNextLine || lineNumber + 1 != errmsg.lineNumber)
                 return Result::None;
         }
-        if (!fileName.empty() && !PathMatch::match(fileName, errmsg.getFileName()))
+        if (!fileName.empty() && !isFileNameMatch(errmsg.getFileName()))
             return Result::None;
         if (hash > 0 && hash != errmsg.hash)
             return Result::Checked;
@@ -456,6 +456,21 @@ SuppressionList::Suppression::Result SuppressionList::Suppression::isSuppressed(
     return Result::Matched;
 }
 
+bool SuppressionList::Suppression::isFileNameMatch(const std::string &errorFileName) const
+{
+    const auto it = mFileNameMatchCache.find(errorFileName);
+    if (it != mFileNameMatchCache.end())
+        return it->second;
+
+    const bool result = PathMatch::match(fileName, errorFileName);
+
+    if (mFileNameMatchCache.size() >= mFileNameMatchCacheMaxEntries)
+        mFileNameMatchCache.clear();
+
+    mFileNameMatchCache.emplace(errorFileName, result);
+    return result;
+}
+
 bool SuppressionList::Suppression::isMatch(const SuppressionList::ErrorMessage &errmsg)
 {
     switch (isSuppressed(errmsg)) {

lib/suppressions.h

@@ -168,6 +168,12 @@ class CPPCHECKLIB SuppressionList {
         bool isPolyspace{};
 
         enum : std::int8_t { NO_LINE = -1 };
+
+    private:
+        bool isFileNameMatch(const std::string &errorFileName) const;
+
+        static constexpr std::size_t mFileNameMatchCacheMaxEntries = 256;
+        mutable std::map<std::string, bool> mFileNameMatchCache;
     };
 
     /**

test/cli/performance_test.py

@@ -438,3 +438,43 @@ def test_slow_many_headers(tmpdir):
     cppcheck(['-DCONFIG0', c_file])
     end = time.perf_counter_ns()
     assert end - start < 2 * 10**9 # max 2 sec
+
+
+@pytest.mark.skipif(sys.platform == 'darwin', reason='GitHub macOS runners are too slow')
+@pytest.mark.timeout(10)
+def test_large_number_of_violations_and_suppressions(tmpdir):
+    filename_main = os.path.join(tmpdir, 'main.c')
+    # This name causes the PathMatch::match() to iterate ~70 times, which is not unrealistic for a header file placed in subdirs.
+    # The number of iterations also depends on how the suppressions are written.
+    header_name = 'long_filename_to_simulate_a_more_realistic_header_placed_in_subdirs.h'
+    header_file = os.path.join(tmpdir, header_name)
+    suppressions_file = os.path.join(tmpdir, 'suppressions.txt')
+
+    # Create a main file that includes a header and returns 0.
+    with open(filename_main, "w") as f:
+        f.write(f'#include "{header_name}"\n\n')
+        f.write('int main() \n')
+        f.write('{\n')
+        f.write('    return 0;\n')
+        f.write('}\n')
+
+    # Create a header file with macro definitions that violates misra because they start with an underscore.
+    with open(header_file, "w") as f:
+        f.write(f'#ifndef {header_name.upper().replace(".", "_")}\n')
+        f.write(f'#define {header_name.upper().replace(".", "_")}\n')
+        f.write('\n')
+        for i in range(5000):
+            f.write(f'#define _FOO{i} {i}\n')
+        f.write('\n')
+        f.write(f'#endif // {header_name.upper().replace(".", "_")}\n')
+
+    # Create a suppressions file that suppresses the misra violation for the macros in the header file
+    with open(suppressions_file, "w") as f:
+        f.write(f'misra-c2012-2.5:{header_name}\n')
+        f.write(f'misra-c2012-21.1:{header_name}\n')
+        # Create other suppressions that don't match the file name to test that iterating the suppressions are faster with cache
+        for i in range(300):
+            f.write(f'misra-c2012-2.5:**/{i}/{header_name}\n')
+            f.write(f'misra-c2012-21.1:**/{i}/{header_name}\n')
+
+    cppcheck([filename_main] + ['--addon=misra.py', '--check-level=exhaustive', '--enable=all', f'--suppressions-list={suppressions_file}'])