regexp -indices should return character indices

Not byte indices

Reported-by: dbohdan <dbohdan@dbohdan.com>
Signed-off-by: Steve Bennett <steveb@workware.net.au>

2 files changed, 11 insertions, 8 deletions

diff --git a/jim-regexp.c b/jim-regexp.c
index 81f3207..3134598 100644
--- a/jim-regexp.c
+++ b/jim-regexp.c
@@ -281,16 +281,15 @@ int Jim_RegexpCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
             }
         }
         else {
-            int len = pmatch[j].rm_eo - pmatch[j].rm_so;
-
             if (opt_indices) {
-                Jim_ListAppendElement(interp, resultObj, Jim_NewIntObj(interp,
-                        offset + pmatch[j].rm_so));
-                Jim_ListAppendElement(interp, resultObj, Jim_NewIntObj(interp,
-                        offset + pmatch[j].rm_so + len - 1));
+                /* rm_so and rm_eo are byte offsets. We need char offsets */
+                int so = utf8_strlen(source_str, pmatch[j].rm_so);
+                int eo = utf8_strlen(source_str + pmatch[j].rm_so, pmatch[j].rm_eo);
+                Jim_ListAppendElement(interp, resultObj, Jim_NewIntObj(interp, offset + so));
+                Jim_ListAppendElement(interp, resultObj, Jim_NewIntObj(interp, offset + eo - 1));
             }
             else {
-                Jim_AppendString(interp, resultObj, source_str + pmatch[j].rm_so, len);
+                Jim_AppendString(interp, resultObj, source_str + pmatch[j].rm_so, pmatch[j].rm_eo - pmatch[j].rm_so);
             }
         }
 
@@ -311,7 +310,7 @@ int Jim_RegexpCmd(Jim_Interp *interp, int argc, Jim_Obj *const *argv)
   try_next_match:
     if (opt_all && (pattern[0] != '^' || (regcomp_flags & REG_NEWLINE)) && *source_str) {
         if (pmatch[0].rm_eo) {
-            offset += pmatch[0].rm_eo;
+            offset += utf8_strlen(source_str, pmatch[0].rm_eo);
             source_str += pmatch[0].rm_eo;
         }
         else {
diff --git a/tests/regexp2.test b/tests/regexp2.test
index 76735e6..1aee8cd 100644
--- a/tests/regexp2.test
+++ b/tests/regexp2.test
@@ -633,6 +633,7 @@ test regexpComp-16.5 {regexp -start with utf8} utf8 {
 test regexpComp-16.6 {regexp -start with utf8} utf8 {
     regsub -start 1 . \u0442\u0435\u0441\u0442 x
 } \u0442x\u0441\u0442
+
 test regexpComp-17.1 {regexp -inline} {
     regexp -inline b ababa
 } {b}
@@ -654,6 +655,9 @@ test regexpComp-17.6 {regexp -inline no matches} {
 test regexpComp-17.7 {regexp -inline, no matchvars allowed} {
     list [catch {regexp -inline b abc match} msg] $msg
 } {1 {regexp match variables not allowed when using -inline}}
+test regexpComp-17.8 {regexp -indices utf8} utf8 {
+    regexp -all -inline -start 1 -indices . \u0442\u0435\u0441\u0442
+} {{1 1} {2 2} {3 3}}
 
 test regexpComp-18.1 {regexp -all} {
     regexp -all b bbbbb