From e90b85cc98a24003f2d673bab2c255ab3dce66e7 Mon Sep 17 00:00:00 2001
From: Jeff Epler <jepler@gmail.com>
Date: Mon, 5 Sep 2022 07:58:04 -0500
Subject: [PATCH] extmod/modure: Convert byte offsets to unicode indices when
 necessary.

And add a test.

Fixes issue #9202.

Signed-off-by: Jeff Epler <jepler@gmail.com>
---
 extmod/modure.c              | 16 ++++++++++++++++
 tests/unicode/unicode_ure.py | 32 ++++++++++++++++++++++++++++++++
 2 files changed, 48 insertions(+)
 create mode 100644 tests/unicode/unicode_ure.py

diff --git a/extmod/modure.c b/extmod/modure.c
index 799fef13b1..a674d66499 100644
--- a/extmod/modure.c
+++ b/extmod/modure.c
@@ -33,6 +33,10 @@
 #include "py/objstr.h"
 #include "py/stackctrl.h"
 
+#if MICROPY_PY_BUILTINS_STR_UNICODE
+#include "py/unicode.h"
+#endif
+
 #if MICROPY_PY_URE
 
 #define re1_5_stack_chk() MP_STACK_CHECK()
@@ -121,6 +125,18 @@ STATIC void match_span_helper(size_t n_args, const mp_obj_t *args, mp_obj_t span
         e = self->caps[no * 2 + 1] - begin;
     }
 
+    #if MICROPY_PY_BUILTINS_STR_UNICODE
+    if (mp_obj_get_type(self->str) == &mp_type_str) {
+        const byte *begin = (const byte *)mp_obj_str_get_str(self->str);
+        if (s != -1) {
+            s = utf8_ptr_to_index(begin, begin + s);
+        }
+        if (e != -1) {
+            e = utf8_ptr_to_index(begin, begin + e);
+        }
+    }
+    #endif
+
     span[0] = mp_obj_new_int(s);
     span[1] = mp_obj_new_int(e);
 }
diff --git a/tests/unicode/unicode_ure.py b/tests/unicode/unicode_ure.py
new file mode 100644
index 0000000000..5a5dc60054
--- /dev/null
+++ b/tests/unicode/unicode_ure.py
@@ -0,0 +1,32 @@
+# test match.span() for unicode strings
+
+try:
+    import ure as re
+except ImportError:
+    try:
+        import re
+    except ImportError:
+        print("SKIP")
+        raise SystemExit
+
+try:
+    m = re.match(".", "a")
+    m.span
+except AttributeError:
+    print("SKIP")
+    raise SystemExit
+
+
+def print_spans(match):
+    print("----")
+    try:
+        i = 0
+        while True:
+            print(match.span(i), match.start(i), match.end(i))
+            i += 1
+    except IndexError:
+        pass
+
+
+m = re.match(r"([0-9]*)(([a-z]*)([0-9]*))", "1234\u2764567")
+print_spans(m)