[Bug 3466099] BOM in Unicode

FossilOrigin-Name: e17fc0e07b10b56f70e42164b20cbb4c570555b8

4 files changed, 41 insertions, 6 deletions

diff --git a/ChangeLog b/ChangeLog
index 0dbd68c..90f1e18 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2012-02-29  Jan Nijtmans  <nijtmans@users.sf.net>
+
+	* generic/tclIOUtil.c:	[Bug 3466099] BOM in Unicode
+	* generic/tclEncoding.c:
+	* tests/source.test
+
 2012-02-09  Don Porter  <dgp@users.sourceforge.net>
 
 	* generic/tclStringObj.c:	[Bug 3484402] Correct Off-By-One
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) {
diff --git a/tests/source.test b/tests/source.test
index 3a1454c..1e5b732 100644
--- a/tests/source.test
+++ b/tests/source.test
@@ -118,6 +118,21 @@ test source-2.6 {source error conditions} -setup {
 	{couldn't read file "*_non_existent_": no such file or directory} \
 	{POSIX ENOENT {no such file or directory}}]
 
+test source-2.7 {utf-8 with BOM} -setup {
+    set sourcefile [makeFile {} source.file]
+    set saveencoding [encoding system]
+} -body {
+    encoding system utf-8
+    set out [open $sourcefile w]
+    puts $out "\ufffeset y new-y"
+    close $out
+    set y old-y
+    source $sourcefile
+    return $y
+} -cleanup {
+    removeFile $sourcefile
+    encoding system $saveencoding
+} -result {new-y}
 
 test source-3.1 {return in middle of source file} -setup {
     set sourcefile [makeFile {