Issue #22437: Number of capturing groups in regular expression is no longer

limited by 100.

2 files changed, 40 insertions, 18 deletions

diff --git a/Modules/_sre.c b/Modules/_sre.c
index 13479ba..5c3d105 100644
--- a/Modules/_sre.c
+++ b/Modules/_sre.c
@@ -357,6 +357,11 @@ state_init(SRE_STATE* state, PatternObject* pattern, PyObject* string,
 
     memset(state, 0, sizeof(SRE_STATE));
 
+    state->mark = PyMem_New(void *, pattern->groups * 2);
+    if (!state->mark) {
+        PyErr_NoMemory();
+        goto err;
+    }
     state->lastmark = -1;
     state->lastindex = -1;
 
@@ -409,6 +414,8 @@ state_init(SRE_STATE* state, PatternObject* pattern, PyObject* string,
 
     return string;
   err:
+    PyMem_Del(state->mark);
+    state->mark = NULL;
     if (state->buffer.buf)
         PyBuffer_Release(&state->buffer);
     return NULL;
@@ -421,6 +428,8 @@ state_fini(SRE_STATE* state)
         PyBuffer_Release(&state->buffer);
     Py_XDECREF(state->string);
     data_stack_dealloc(state);
+    PyMem_Del(state->mark);
+    state->mark = NULL;
 }
 
 /* calculate offset from start of string */
@@ -560,6 +569,7 @@ pattern_match(PatternObject *self, PyObject *args, PyObject *kwargs)
     PyObject *pattern = NULL;
     SRE_STATE state;
     Py_ssize_t status;
+    PyObject *match;
 
     if (!PyArg_ParseTupleAndKeywords(args, kwargs,
         "|Onn$O:match", _keywords,
@@ -579,12 +589,14 @@ pattern_match(PatternObject *self, PyObject *args, PyObject *kwargs)
     status = sre_match(&state, PatternObject_GetCode(self), 0);
 
     TRACE(("|%p|%p|END\n", PatternObject_GetCode(self), state.ptr));
-    if (PyErr_Occurred())
+    if (PyErr_Occurred()) {
+        state_fini(&state);
         return NULL;
+    }
 
+    match = pattern_new_match(self, &state, status);
     state_fini(&state);
-
-    return (PyObject *)pattern_new_match(self, &state, status);
+    return match;
 }
 
 static PyObject*
@@ -592,6 +604,7 @@ pattern_fullmatch(PatternObject* self, PyObject* args, PyObject* kw)
 {
     SRE_STATE state;
     Py_ssize_t status;
+    PyObject *match;
 
     PyObject *string = NULL, *string2 = NULL;
     Py_ssize_t start = 0;
@@ -616,12 +629,14 @@ pattern_fullmatch(PatternObject* self, PyObject* args, PyObject* kw)
     status = sre_match(&state, PatternObject_GetCode(self), 1);
 
     TRACE(("|%p|%p|END\n", PatternObject_GetCode(self), state.ptr));
-    if (PyErr_Occurred())
+    if (PyErr_Occurred()) {
+        state_fini(&state);
         return NULL;
+    }
 
+    match = pattern_new_match(self, &state, status);
     state_fini(&state);
-
-    return pattern_new_match(self, &state, status);
+    return match;
 }
 
 static PyObject*
@@ -629,6 +644,7 @@ pattern_search(PatternObject* self, PyObject* args, PyObject* kw)
 {
     SRE_STATE state;
     Py_ssize_t status;
+    PyObject *match;
 
     PyObject *string = NULL, *string2 = NULL;
     Py_ssize_t start = 0;
@@ -652,12 +668,14 @@ pattern_search(PatternObject* self, PyObject* args, PyObject* kw)
 
     TRACE(("|%p|%p|END\n", PatternObject_GetCode(self), state.ptr));
 
-    state_fini(&state);
-
-    if (PyErr_Occurred())
+    if (PyErr_Occurred()) {
+        state_fini(&state);
         return NULL;
+    }
 
-    return pattern_new_match(self, &state, status);
+    match = pattern_new_match(self, &state, status);
+    state_fini(&state);
+    return match;
 }
 
 static PyObject*
@@ -1417,7 +1435,7 @@ _compile(PyObject* self_, PyObject* args)
     PyObject* groupindex = NULL;
     PyObject* indexgroup = NULL;
 
-    if (!PyArg_ParseTuple(args, "OiO!|nOO", &pattern, &flags,
+    if (!PyArg_ParseTuple(args, "OiO!nOO", &pattern, &flags,
                           &PyList_Type, &code, &groups,
                           &groupindex, &indexgroup))
         return NULL;
@@ -1933,10 +1951,9 @@ _validate_inner(SRE_CODE *code, SRE_CODE *end, Py_ssize_t groups)
 static int
 _validate_outer(SRE_CODE *code, SRE_CODE *end, Py_ssize_t groups)
 {
-    if (groups < 0 || groups > 100 || code >= end || end[-1] != SRE_OP_SUCCESS)
+    if (groups < 0 || (size_t)groups > SRE_MAXGROUPS ||
+        code >= end || end[-1] != SRE_OP_SUCCESS)
         FAIL;
-    if (groups == 0)  /* fix for simplejson */
-        groups = 100; /* 100 groups should always be safe */
     return _validate_inner(code, end-1, groups);
 }
 
@@ -2747,6 +2764,12 @@ PyMODINIT_FUNC PyInit__sre(void)
         Py_DECREF(x);
     }
 
+    x = PyLong_FromUnsignedLong(SRE_MAXGROUPS);
+    if (x) {
+        PyDict_SetItemString(d, "MAXGROUPS", x);
+        Py_DECREF(x);
+    }
+
     x = PyUnicode_FromString(copyright);
     if (x) {
         PyDict_SetItemString(d, "copyright", x);
diff --git a/Modules/sre.h b/Modules/sre.h
index 42fe28d..35d198f 100644
--- a/Modules/sre.h
+++ b/Modules/sre.h
@@ -18,8 +18,10 @@
 #define SRE_CODE Py_UCS4
 #if SIZEOF_SIZE_T > 4
 # define SRE_MAXREPEAT (~(SRE_CODE)0)
+# define SRE_MAXGROUPS ((~(SRE_CODE)0) / 2)
 #else
 # define SRE_MAXREPEAT ((SRE_CODE)PY_SSIZE_T_MAX)
+# define SRE_MAXGROUPS ((SRE_CODE)PY_SSIZE_T_MAX / SIZEOF_SIZE_T / 2)
 #endif
 
 typedef struct {
@@ -52,9 +54,6 @@ typedef struct {
 
 typedef unsigned int (*SRE_TOLOWER_HOOK)(unsigned int ch);
 
-/* FIXME: <fl> shouldn't be a constant, really... */
-#define SRE_MARK_SIZE 200
-
 typedef struct SRE_REPEAT_T {
     Py_ssize_t count;
     SRE_CODE* pattern; /* points to REPEAT operator arguments */
@@ -76,7 +75,7 @@ typedef struct {
     /* registers */
     Py_ssize_t lastindex;
     Py_ssize_t lastmark;
-    void* mark[SRE_MARK_SIZE];
+    void** mark;
     /* dynamically allocated stuff */
     char* data_stack;
     size_t data_stack_size;