Fix "string tolower" and friends for handling unpaired surrogates correctly. Also add test-cases for those situations.

Various typo's in comments.

1 files changed, 15 insertions, 0 deletions

diff --git a/tests/utf.test b/tests/utf.test
index 67a6778..e820359 100644
--- a/tests/utf.test
+++ b/tests/utf.test
@@ -158,6 +158,12 @@ test utf-8.3 {Tcl_UniCharAtIndex: index > 0} {
 test utf-8.4 {Tcl_UniCharAtIndex: index > 0} {
     string index \u4e4e\u25a\xff\u543 2
 } "\uff"
+test utf-8.5 {Tcl_UniCharAtIndex: upper surrogate} {
+    string index \ud842 0
+} "\ud842"
+test utf-8.5 {Tcl_UniCharAtIndex: lower surrogate} {
+    string index \udc42 0
+} "\udc42"
 
 test utf-9.1 {Tcl_UtfAtIndex: index = 0} {
     string range abcd 0 2
@@ -263,6 +269,9 @@ test utf-11.4 {Tcl_UtfToUpper} {
 test utf-11.5 {Tcl_UtfToUpper Georgian (new in Unicode 11)} {
     string toupper \u10d0\u1c90
 } \u1c90\u1c90
+test utf-11.6 {Tcl_UtfToUpper low/high surrogate)} {
+    string toupper \udc24\ud824
+} \udc24\ud824
 
 test utf-12.1 {Tcl_UtfToLower} {
     string tolower {}
@@ -279,6 +288,9 @@ test utf-12.4 {Tcl_UtfToLower} {
 test utf-12.5 {Tcl_UtfToLower Georgian (new in Unicode 11)} {
     string tolower \u10d0\u1c90
 } \u10d0\u10d0
+test utf-12.6 {Tcl_UtfToUpper low/high surrogate)} {
+    string tolower \udc24\ud824
+} \udc24\ud824
 
 test utf-13.1 {Tcl_UtfToTitle} {
     string totitle {}
@@ -298,6 +310,9 @@ test utf-13.5 {Tcl_UtfToTitle Georgian (new in Unicode 11)} {
 test utf-13.6 {Tcl_UtfToTitle Georgian (new in Unicode 11)} {
     string totitle \u1c90\u10d0
 } \u1c90\u10d0
+test utf-13.7 {Tcl_UtfToTitle low/high surrogate)} {
+    string totitle \udc24\ud824
+} \udc24\ud824
 
 test utf-14.1 {Tcl_UtfNcasecmp} {
     string compare -nocase a b