StdLib/LibC/Locale/multibyte_Utf8.c: Fix obscure corner cases in wide to multibyte and multibyte to wide character conversions. The majority of problems center around the interpretation of the Length or Limit parameter when the Destination parameter is NULL.

DecodeOneStateful: Properly handle combinations of Src, Dest, or Len being NULL or 0.

EncodeUtf8: Do not zero-terminate the result string in this worker function.

mbsrtowcs: Remove test for **src == '\0', as per ISO/IEC 9899:199409.  Allows "".

wcsrtombs:  The C Language standard, ISO/IEC 9899:199409, states that the wcsrtombs() function will stop before encountering the terminating NUL character only if Dest is NOT NULL.  This implies that if Dest is NULL, the Limit parameter will be ignored.  In order to avoid system hangs, if Dest is NULL a Limit value of ASCII_STRING_MAX is automatically used.  Also fixed a typo in the function header comment.

With these changes, StdLib now passes all of the C Language Standards Compliance Tests for ISO/IEC 9899:199409 (C95).

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by:  daryl.mcdaniel@intel.com
Reviewed-by:    erik.c.bjorge@intel.com


git-svn-id: https://edk2.svn.sourceforge.net/svnroot/edk2/trunk/edk2@14358 6f19259b-4bc3-4df7-8a09-765794883524

1 files changed, 22 insertions, 24 deletions

diff --git a/StdLib/LibC/Locale/multibyte_Utf8.c b/StdLib/LibC/Locale/multibyte_Utf8.c
index ec9b01265..ffe3dee23 100644
--- a/StdLib/LibC/Locale/multibyte_Utf8.c
+++ b/StdLib/LibC/Locale/multibyte_Utf8.c
@@ -197,19 +197,24 @@ DecodeOneStateful(
   int           NumConv;

   unsigned char ch;

 

-  if((Src == NULL) || (*Src == '\0')) {

-    return 0;

-  }

   if(pS == NULL) {

     pS = &LocalConvState;

   }

-  SrcEnd  = Src + Len;

   NumConv = 0;

-  while(Src < SrcEnd) {

-    ch = (unsigned char)*Src++;

-    NumConv = ProcessOneByte(ch, pS);

-    if(NumConv != -2)

-      break;

+  if(Src != NULL) {

+    if(*Src != 0) {

+      SrcEnd  = Src + Len;

+      while(Src < SrcEnd) {

+        ch = (unsigned char)*Src++;

+        NumConv = ProcessOneByte(ch, pS);

+        if(NumConv != -2) {

+          break;

+        }

+      }

+    }

+    else if(Dest != NULL) {

+      *Dest = 0;

+    }

   }

   if((NumConv > 0) && (Dest != NULL)) {

     Dest[0] = pS->D[0];

@@ -416,14 +421,6 @@ EncodeUtf8(char *Dest, wchar_t ch)
   */

   if(Dest != NULL) {        // Save character if Dest is not NULL

     memcpy(Dest, Buff, NumInBuff);

-

-    if(ch != 0) {

-      // Terminate the destination string.

-      Dest[NumInBuff] = '\0';

-    }

-    else {

-      NumInBuff = 0;

-    }

   }

   return NumInBuff;             // Tell the caller

 }

@@ -646,7 +643,7 @@ mbsrtowcs(
   size_t        RetVal = 0;

   const char   *MySrc;

 

-  if((src == NULL) || (*src == NULL) || (**src == '\0')) {

+  if((src == NULL) || (*src == NULL)) {

     return 0;

   }

 

@@ -855,7 +852,7 @@ wctomb(
 }

 

 /** The wcsrtombs function converts a sequence of wide characters from the array

-    indirectly pointed to by Dest into a sequence of corresponding multibyte

+    indirectly pointed to by Src into a sequence of corresponding multibyte

     characters that begins in the conversion state described by the object

     pointed to by ps.

 

@@ -914,15 +911,16 @@ wcsrtombs(
     return (0);

 

   if (Dest == NULL) {

-    if(MaxBytes <= 0) {

-      MaxBytes = ASCII_STRING_MAX;

-    }

-    NumStored = EstimateWtoM(*Src, MaxBytes, NULL);

+    NumStored = EstimateWtoM(*Src, ASCII_STRING_MAX, NULL);

   }

   else {

-    while (OneWcToMcLen(InCh = *(*Src)++) <= MaxBytes) {

+    if((MaxBytes < 0) || (MaxBytes > ASCII_STRING_MAX)) {

+      MaxBytes = ASCII_STRING_MAX;

+    }

+    while ((MaxBytes > 0) && (OneWcToMcLen(InCh = *(*Src)++) <= MaxBytes)) {

       if(InCh == 0) {

         *Src = NULL;

+        *Dest = 0;      // NUL terminate Dest string, but don't count the NUL

         break;

       }

       count = (int)wcrtomb(Dest, InCh, NULL);