FIX: Decode Raw UTF-16 data from Conn.getinfo()

Author: bewithgaurav

mssql_python/connection.py

@@ -1198,21 +1198,21 @@ def getinfo(self, info_type: int) -> Union[str, int, bool, None]:
                     # Make sure we use the correct amount of data based on length
                     actual_data = data[:length]
 
-                    # Now decode the string data
-                    try:
-                        return actual_data.decode("utf-8").rstrip("\0")
-                    except UnicodeDecodeError:
+                    # SQLGetInfoW returns UTF-16LE encoded strings
+                    # Try encodings in order: UTF-16LE (Windows), UTF-8, Latin-1
+                    for encoding in ("utf-16-le", "utf-8", "latin1"):
                         try:
-                            return actual_data.decode("latin1").rstrip("\0")
-                        except Exception as e:
-                            logger.debug(
-                                "error",
-                                "Failed to decode string in getinfo: %s. "
-                                "Returning None to avoid silent corruption.",
-                                e,
-                            )
-                            # Explicitly return None to signal decoding failure
-                            return None
+                            return actual_data.decode(encoding).rstrip("\0")
+                        except UnicodeDecodeError:
+                            continue
+                    
+                    # All decodings failed
+                    logger.debug(
+                        "error",
+                        "Failed to decode string in getinfo with any supported encoding. "
+                        "Returning None to avoid silent corruption.",
+                    )
+                    return None
                 else:
                     # If it's not bytes, return as is
                     return data

mssql_python/pybind/build.bat

@@ -1,10 +1,30 @@
 @echo off
 setlocal enabledelayedexpansion
 
-REM Usage: build.bat [ARCH], If ARCH is not specified, it defaults to x64.
+REM Usage: build.bat [ARCH] [MODE], If ARCH is not specified, it defaults to x64.
+REM MODE can be 'profile' or '--profile' to enable profiling instrumentation
 set ARCH=%1
 if "%ARCH%"=="" set ARCH=x64
+
+REM Check for profiling mode
+set PROFILING_MODE=0
+if /i "%2"=="profile" set PROFILING_MODE=1
+if /i "%2"=="--profile" set PROFILING_MODE=1
+if /i "%1"=="profile" (
+    set PROFILING_MODE=1
+    set ARCH=x64
+)
+if /i "%1"=="--profile" (
+    set PROFILING_MODE=1
+    set ARCH=x64
+)
+
 echo [DIAGNOSTIC] Target Architecture set to: %ARCH%
+if %PROFILING_MODE%==1 (
+    echo [MODE] C++ Profiling: ENABLED
+) else (
+    echo [MODE] C++ Profiling: DISABLED
+)
 
 REM Clean up main build directory if it exists
 echo Checking for main build directory...
@@ -109,8 +129,13 @@ if errorlevel 1 (
 )
 
 REM Now invoke CMake with correct source path (options first, path last!)
-echo [DIAGNOSTIC] Running CMake configure with: cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% "%SOURCE_DIR:~0,-1%"
-cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% "%SOURCE_DIR:~0,-1%"
+if %PROFILING_MODE%==1 (
+    echo [DIAGNOSTIC] Running CMake configure with profiling: cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% -DCMAKE_CXX_FLAGS="/DENABLE_PROFILING" -DCMAKE_C_FLAGS="/DENABLE_PROFILING" "%SOURCE_DIR:~0,-1%"
+    cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% -DCMAKE_CXX_FLAGS="/DENABLE_PROFILING" -DCMAKE_C_FLAGS="/DENABLE_PROFILING" "%SOURCE_DIR:~0,-1%"
+) else (
+    echo [DIAGNOSTIC] Running CMake configure with: cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% "%SOURCE_DIR:~0,-1%"
+    cmake -A %PLATFORM_NAME% -DARCHITECTURE=%ARCH% "%SOURCE_DIR:~0,-1%"
+)
 echo [DIAGNOSTIC] CMake configure exit code: %errorlevel%
 if errorlevel 1 (
     echo [ERROR] CMake configuration failed
@@ -157,22 +182,6 @@ if exist "%OUTPUT_DIR%\%PYD_NAME%" (
         echo [WARNING] PDB file !PDB_NAME! not found in output directory.
     )
 
-    setlocal enabledelayedexpansion
-    for %%I in ("%SOURCE_DIR%..") do (
-        set PARENT_DIR=%%~fI
-    )
-    echo [DIAGNOSTIC] Parent is: !PARENT_DIR!
-
-    set VCREDIST_DLL_PATH=!PARENT_DIR!\libs\windows\!ARCH!\vcredist\msvcp140.dll
-    echo [DIAGNOSTIC] Looking for msvcp140.dll at "!VCREDIST_DLL_PATH!"
-
-    if exist "!VCREDIST_DLL_PATH!" (
-        copy /Y "!VCREDIST_DLL_PATH!" "%SOURCE_DIR%\.."
-        echo [SUCCESS] Copied msvcp140.dll from !VCREDIST_DLL_PATH! to "%SOURCE_DIR%\.."
-    ) else (
-        echo [ERROR] Could not find msvcp140.dll at "!VCREDIST_DLL_PATH!"
-        exit /b 1
-    )
 ) else (
     echo [ERROR] Could not find built .pyd file: %PYD_NAME%
     REM Exit with an error code here if the .pyd file is not found

tests/test_003_connection.py

@@ -5471,6 +5471,33 @@ def test_getinfo_basic_driver_info(db_connection):
         pytest.fail(f"getinfo failed for basic driver info: {e}")
 
 
+def test_getinfo_string_encoding_utf16(db_connection):
+    """Test that string values from getinfo are properly decoded from UTF-16."""
+    
+    # Test string info types that should not contain null bytes
+    string_info_types = [
+        ("SQL_DRIVER_VER", sql_const.SQL_DRIVER_VER.value),
+        ("SQL_DRIVER_NAME", sql_const.SQL_DRIVER_NAME.value),
+        ("SQL_DRIVER_ODBC_VER", sql_const.SQL_DRIVER_ODBC_VER.value),
+        ("SQL_SERVER_NAME", sql_const.SQL_SERVER_NAME.value),
+    ]
+    
+    for name, info_type in string_info_types:
+        result = db_connection.getinfo(info_type)
+        
+        if result is not None:
+            # Verify it's a string
+            assert isinstance(result, str), \
+                f"{name}: Expected str, got {type(result).__name__}"
+            
+            # Verify no null bytes (indicates UTF-16 decoded as UTF-8 bug)
+            assert '\x00' not in result, \
+                f"{name} contains null bytes, likely UTF-16/UTF-8 encoding mismatch: {repr(result)}"
+            
+            # Verify it's not empty (optional, but good sanity check)
+            assert len(result) > 0, f"{name} returned empty string"
+
+
 def test_getinfo_sql_support(db_connection):
     """Test SQL support and conformance info types."""