[Bug 3466099] BOM in Unicode

2 files changed, 20 insertions, 6 deletions

diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c
index 9432c05..a4a7027 100644
--- a/generic/tclEncoding.c
+++ b/generic/tclEncoding.c
@@ -764,13 +764,13 @@ Tcl_GetEncodingNames(interp)
 int
 Tcl_SetSystemEncoding(interp, name)
     Tcl_Interp *interp;		/* Interp for error reporting, if not NULL. */
-    CONST char *name;		/* The name of the desired encoding, or NULL
+    CONST char *name;		/* The name of the desired encoding, or NULL/""
 				 * to reset to default encoding. */
 {
     Tcl_Encoding encoding;
     Encoding *encodingPtr;
 
-    if (name == NULL) {
+    if (!name || !*name) {
 	Tcl_MutexLock(&encodingMutex);
 	encoding = defaultEncoding;
 	encodingPtr = (Encoding *) encoding;
diff --git a/generic/tclIOUtil.c b/generic/tclIOUtil.c
index b54b76b..5a8d022 100644
--- a/generic/tclIOUtil.c
+++ b/generic/tclIOUtil.c
@@ -1755,11 +1755,25 @@ Tcl_FSEvalFile(interp, pathPtr)
      * [Bug: 2040]
      */
     Tcl_SetChannelOption(interp, chan, "-eofchar", "\32");
-    if (Tcl_ReadChars(chan, objPtr, -1, 0) < 0) {
-        Tcl_Close(interp, chan);
+    /* Try to read first character of stream, so we can
+     * check for utf-8 BOM to be handled especially.
+     */
+    if (Tcl_ReadChars(chan, objPtr, 1, 0) < 0) {
+	Tcl_Close(interp, chan);
 	Tcl_AppendResult(interp, "couldn't read file \"", 
-		Tcl_GetString(pathPtr),
-		"\": ", Tcl_PosixError(interp), (char *) NULL);
+		Tcl_GetString(pathPtr),	"\": ", Tcl_PosixError(interp), NULL);
+	goto end;
+    }
+    string = Tcl_GetString(objPtr);
+    /*
+     * If first character is not a BOM, append the remaining characters,
+     * otherwise replace them [Bug 3466099].
+     */
+    if (Tcl_ReadChars(chan, objPtr, -1,
+	    memcmp(string, "\xef\xbf\xbe", 3)) < 0) {
+	Tcl_Close(interp, chan);
+	Tcl_AppendResult(interp, "couldn't read file \"",
+		Tcl_GetString(pathPtr), "\": ", Tcl_PosixError(interp), NULL);
 	goto end;
     }
     if (Tcl_Close(interp, chan) != TCL_OK) {