Implement TIP 318.

3 files changed, 34 insertions, 8 deletions

diff --git a/ChangeLog b/ChangeLog
index f6b346e..e3c731c 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,12 @@
+2008-09-29  Donal K. Fellows  <dkf@users.sf.net>
+
+	TIP #318 IMPLEMENTATION
+
+	* generic/tclCmdMZ.c (StringTrimCmd,StringTrimLCmd,StringTrimRCmd):
+	Update the default set of trimmed characters to include some from
+	the larger UNICODE space. Factor out the default trim set into a
+	macro so that it is easier to keep them in synch.
+
 2008-09-28  Donal K. Fellows  <dkf@users.sf.net>
 
 	TIP #314 IMPLEMENTATION
diff --git a/generic/tclCmdMZ.c b/generic/tclCmdMZ.c
index 227e7b8..dfeb59c 100644
--- a/generic/tclCmdMZ.c
+++ b/generic/tclCmdMZ.c
@@ -15,7 +15,7 @@
  * See the file "license.terms" for information on usage and redistribution of
  * this file, and for a DISCLAIMER OF ALL WARRANTIES.
  *
- * RCS: @(#) $Id: tclCmdMZ.c,v 1.169 2008/07/31 20:01:40 msofer Exp $
+ * RCS: @(#) $Id: tclCmdMZ.c,v 1.170 2008/09/29 08:20:34 dkf Exp $
  */
 
 #include "tclInt.h"
@@ -23,6 +23,14 @@
 
 static int		UniCharIsAscii(int character);
 
+/*
+ * Default set of characters to trim in [string trim] and friends. This is a
+ * UTF-8 literal string containing space, tab, newline, carriage return,
+ * ethiopic wordspace (U+1361), ogham space mark (U+1680), and ideographic
+ * space (U+3000). [TIP #318]
+ */
+
+#define DEFAULT_TRIM_SET " \t\n\r\xe1\x8d\xa1\xe1\x9a\x80\xe3\x80\x80"
 
 /*
  *----------------------------------------------------------------------
@@ -3069,8 +3077,8 @@ StringTrimCmd(
     if (objc == 3) {
 	string2 = TclGetStringFromObj(objv[2], &length2);
     } else if (objc == 2) {
-	string2 = " \t\n\r";
-	length2 = strlen(string2);
+	string2 = DEFAULT_TRIM_SET;
+	length2 = strlen(DEFAULT_TRIM_SET);
     } else {
 	Tcl_WrongNumArgs(interp, 1, objv, "string ?chars?");
 	return TCL_ERROR;
@@ -3165,8 +3173,8 @@ StringTrimLCmd(
     if (objc == 3) {
 	string2 = TclGetStringFromObj(objv[2], &length2);
     } else if (objc == 2) {
-	string2 = " \t\n\r";
-	length2 = strlen(string2);
+	string2 = DEFAULT_TRIM_SET;
+	length2 = strlen(DEFAULT_TRIM_SET);
     } else {
 	Tcl_WrongNumArgs(interp, 1, objv, "string ?chars?");
 	return TCL_ERROR;
@@ -3237,8 +3245,8 @@ StringTrimRCmd(
     if (objc == 3) {
 	string2 = TclGetStringFromObj(objv[2], &length2);
     } else if (objc == 2) {
-	string2 = " \t\n\r";
-	length2 = strlen(string2);
+	string2 = DEFAULT_TRIM_SET;
+	length2 = strlen(DEFAULT_TRIM_SET);
     } else {
 	Tcl_WrongNumArgs(interp, 1, objv, "string ?chars?");
 	return TCL_ERROR;
diff --git a/tests/string.test b/tests/string.test
index 64ec56f..c2ddfc8 100644
--- a/tests/string.test
+++ b/tests/string.test
@@ -12,7 +12,7 @@
 # See the file "license.terms" for information on usage and redistribution
 # of this file, and for a DISCLAIMER OF ALL WARRANTIES.
 #
-# RCS: @(#) $Id: string.test,v 1.73 2008/07/19 22:50:38 nijtmans Exp $
+# RCS: @(#) $Id: string.test,v 1.74 2008/09/29 08:20:38 dkf Exp $
 
 if {[lsearch [namespace children] ::tcltest] == -1} {
     package require tcltest
@@ -1407,6 +1407,9 @@ test string-18.10 {string trim} {
 test string-18.11 {string trim, unicode} {
     string trim "\xe7\xe8 AB\xe7C \xe8\xe7" \xe7\xe8
 } " AB\xe7C "
+test string-18.12 {string trim, unicode default} {
+    string trim ABC\u1361\u1680\u3000
+} ABC
 
 test string-19.1 {string trimleft} {
     list [catch {string trimleft} msg] $msg
@@ -1414,6 +1417,9 @@ test string-19.1 {string trimleft} {
 test string-19.2 {string trimleft} {
     string trimleft "    XYZ      "
 } {XYZ      }
+test string-19.3 {string trimleft, unicode default} {
+    string trimleft \u1361\u1680\u3000ABC
+} ABC
 
 test string-20.1 {string trimright errors} {
     list [catch {string trimright} msg] $msg
@@ -1430,6 +1436,9 @@ test string-20.4 {string trimright} {
 test string-20.5 {string trimright} {
     string trimright ""
 } {}
+test string-20.6 {string trimright, unicode default} {
+    string trimright ABC\u1361\u1680\u3000
+} ABC
 
 test string-21.1 {string wordend} {
     list [catch {string wordend a} msg] $msg