Enable unicode when data starts with UTF-8 BOM

Automatically enable the Unicode mode if the data being loaded starts with the UTF-8 BOM.

1 files changed, 14 insertions, 9 deletions

diff --git a/SimpleIni.h b/SimpleIni.h
index 6cf5ced..047c809 100644
--- a/SimpleIni.h
+++ b/SimpleIni.h
@@ -20,7 +20,7 @@
     @section features FEATURES
 
     - MIT Licence allows free use in all software (including GPL and commercial)
-    - multi-platform (Windows 95/98/ME/NT/2K/XP/2003, Windows CE, Linux, Unix)
+    - multi-platform (Windows CE/9x/NT..10/etc, Linux, MacOSX, Unix)
     - loading and saving of INI-style configuration files
     - configuration files can have any newline format on all platforms
     - liberal acceptance of file format
@@ -1414,14 +1414,18 @@ CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadData(
     size_t          a_uDataLen
     )
 {
-    SI_CONVERTER converter(m_bStoreIsUtf8);
-
-    // consume the UTF-8 BOM if it exists
-    if (m_bStoreIsUtf8 && a_uDataLen >= 3) {
-        if (memcmp(a_pData, SI_UTF8_SIGNATURE, 3) == 0) {
-            a_pData    += 3;
-            a_uDataLen -= 3;
-        }
+    if (!a_pData) {
+        return SI_OK;
+    }
+    
+    // if the UTF-8 BOM exists, consume it and set mode to unicode, if we have
+    // already loaded data and try to change mode half-way through then this will
+    // be ignored and we will assert in debug versions
+    if (a_uDataLen >= 3 && memcmp(a_pData, SI_UTF8_SIGNATURE, 3) == 0) {
+        a_pData    += 3;
+        a_uDataLen -= 3;
+        SI_ASSERT(m_bStoreIsUtf8 || !m_pData); // we don't expect mixed mode data
+        SetUnicode();
     }
 
     if (a_uDataLen == 0) {
@@ -1429,6 +1433,7 @@ CSimpleIniTempl<SI_CHAR,SI_STRLESS,SI_CONVERTER>::LoadData(
     }
 
     // determine the length of the converted data
+    SI_CONVERTER converter(m_bStoreIsUtf8);
     size_t uLen = converter.SizeFromStore(a_pData, a_uDataLen);
     if (uLen == (size_t)(-1)) {
         return SI_FAIL;