extmod/modure: Set subject begin_line so ^ doesn't match interior.

Fixes issue #8402. Signed-off-by: Damien George <damien@micropython.org>
micropython · Mar 16, 2022 · 63f0e70 · 63f0e70
1 parent adfd57c
commit 63f0e70
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 3 deletions.
diff --git a/extmod/modure.c b/extmod/modure.c
@@ -187,7 +187,7 @@ STATIC mp_obj_t ure_exec(bool is_anchored, uint n_args, const mp_obj_t *args) {
     }
     Subject subj;
     size_t len;
-    subj.begin = mp_obj_str_get_data(args[1], &len);
+    subj.begin_line = subj.begin = mp_obj_str_get_data(args[1], &len);
     subj.end = subj.begin + len;
     int caps_num = (self->re.sub + 1) * 2;
     mp_obj_match_t *match = m_new_obj_var(mp_obj_match_t, char *, caps_num);
@@ -220,7 +220,7 @@ STATIC mp_obj_t re_split(size_t n_args, const mp_obj_t *args) {
     Subject subj;
     size_t len;
     const mp_obj_type_t *str_type = mp_obj_get_type(args[1]);
-    subj.begin = mp_obj_str_get_data(args[1], &len);
+    subj.begin_line = subj.begin = mp_obj_str_get_data(args[1], &len);
     subj.end = subj.begin + len;
     int caps_num = (self->re.sub + 1) * 2;
 
@@ -280,7 +280,7 @@ STATIC mp_obj_t re_sub_helper(size_t n_args, const mp_obj_t *args) {
     size_t where_len;
     const char *where_str = mp_obj_str_get_data(where, &where_len);
     Subject subj;
-    subj.begin = where_str;
+    subj.begin_line = subj.begin = where_str;
     subj.end = subj.begin + where_len;
     int caps_num = (self->re.sub + 1) * 2;
 

diff --git a/tests/extmod/ure_split.py b/tests/extmod/ure_split.py
@@ -31,3 +31,13 @@
 r = re.compile(b"x")
 s = r.split(b"fooxbar")
 print(s)
+
+# using ^
+r = re.compile("^ab")
+s = r.split("abababcabab")
+print(s)
+
+# using ^ with |
+r = re.compile("^ab|cab")
+s = r.split("abababcabab")
+print(s)
diff --git a/tests/extmod/ure_sub.py b/tests/extmod/ure_sub.py
@@ -75,3 +75,7 @@ def A():
 
 # Include \ in the sub replacement
 print(re.sub("b", "\\\\b", "abc"))
+
+# Using ^, make sure it doesn't repeatedly match
+print(re.sub("^ab", "*", "abababcabab"))
+print(re.sub("^ab|cab", "*", "abababcabab"))