Git issue 546: Partial match not working in some instances with non-greedy capture

Author: Matthew Barnett

changelog.txt

@@ -1,3 +1,7 @@
+Version: 2024.11.6
+
+    Git issue 546: Partial match not working in some instances with non-greedy capture
+
 Version: 2024.9.14
 
     Reverted to actions/download-artifact@v3 and actions/upload-artifact@v3 in main.yml because GitHub Actions failed when using them.

regex_3/_regex.c

@@ -16535,9 +16535,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                           RE_PARTIAL_RIGHT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos >= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search(state, test, pos + 1, limit +
                           length, TRUE, &is_partial);
@@ -16593,9 +16590,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                           RE_PARTIAL_RIGHT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos >= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search_fld(state, test, pos + 1, limit,
                           &new_pos, &is_partial);
@@ -16651,9 +16645,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                         if (pos <= state->text_start && state->partial_side == RE_PARTIAL_LEFT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos <= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search_fld_rev(state, test, pos - 1,
                           limit, &new_pos, &is_partial);
@@ -16713,9 +16704,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                           RE_PARTIAL_RIGHT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos >= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search_ign(state, test, pos + 1, limit +
                           length, TRUE, &is_partial);
@@ -16773,9 +16761,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                         if (pos <= state->text_start && state->partial_side == RE_PARTIAL_LEFT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos <= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search_ign_rev(state, test, pos - 1,
                           limit - length, TRUE, &is_partial);
@@ -16833,9 +16818,6 @@ Py_LOCAL_INLINE(int) basic_match(RE_State* state, BOOL search) {
                         if (pos <= state->text_start && state->partial_side == RE_PARTIAL_LEFT)
                             return RE_ERROR_PARTIAL;
 
-                        if (pos <= limit)
-                            break;
-
                         /* Look for the tail string. */
                         found = string_search_rev(state, test, pos - 1, limit -
                           length, TRUE, &is_partial);

regex_3/regex.py

@@ -241,7 +241,7 @@
   "VERSION1", "X", "VERBOSE", "W", "WORD", "error", "Regex", "__version__",
   "__doc__", "RegexFlag"]
 
-__version__ = "2.5.147"
+__version__ = "2.5.148"
 
 # --------------------------------------------------------------------
 # Public interface.

regex_3/test_regex.py

@@ -4346,6 +4346,16 @@ def test_hg_bugs(self):
         self.assertEqual(regex.match(r"(?i)[^/]*b/xyz", "b/xy", partial=True).span(), (0, 4))
         self.assertEqual(regex.match(r"(?i)[^/]*b/xyz", "b/yz", partial=True), None)
 
+        # Git issue 546: Partial match not working in some instances with non-greedy capture
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>x', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>xyz abc', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>xyz abc foo', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>xyz abc foo ', partial=True)), True)
+        self.assertEqual(bool(regex.match(r'<thinking>.*?</thinking>', '<thinking>xyz abc foo bar', partial=True)), True)
+
     def test_fuzzy_ext(self):
         self.assertEqual(bool(regex.fullmatch(r'(?r)(?:a){e<=1:[a-z]}', 'e')),
           True)

setup.py

@@ -8,7 +8,7 @@
 
 setup(
     name='regex',
-    version='2024.9.11',
+    version='2024.11.6',
     description='Alternative regular expression module, to replace re.',
     long_description=long_description,
     long_description_content_type='text/x-rst',