Cython guard against [c|m|re]alloc failures (pandas-dev#57705)

Author: WillAyd

pandas/_libs/algos.pyx

@@ -180,6 +180,8 @@ def is_lexsorted(list_of_arrays: list) -> bool:
     n = len(list_of_arrays[0])
 
     cdef int64_t **vecs = <int64_t**>malloc(nlevels * sizeof(int64_t*))
+    if vecs is NULL:
+        raise MemoryError()
     for i in range(nlevels):
         arr = list_of_arrays[i]
         assert arr.dtype.name == "int64"

pandas/_libs/groupby.pyx

@@ -81,6 +81,8 @@ cdef float64_t median_linear_mask(float64_t* a, int n, uint8_t* mask) noexcept n
             return NaN
 
         tmp = <float64_t*>malloc((n - na_count) * sizeof(float64_t))
+        if tmp is NULL:
+            raise MemoryError()
 
         j = 0
         for i in range(n):
@@ -118,6 +120,8 @@ cdef float64_t median_linear(float64_t* a, int n) noexcept nogil:
             return NaN
 
         tmp = <float64_t*>malloc((n - na_count) * sizeof(float64_t))
+        if tmp is NULL:
+            raise MemoryError()
 
         j = 0
         for i in range(n):

pandas/_libs/hashing.pyx

@@ -68,7 +68,11 @@ def hash_object_array(
 
     # create an array of bytes
     vecs = <char **>malloc(n * sizeof(char *))
+    if vecs is NULL:
+        raise MemoryError()
     lens = <uint64_t*>malloc(n * sizeof(uint64_t))
+    if lens is NULL:
+        raise MemoryError()
 
     for i in range(n):
         val = arr[i]

pandas/_libs/hashtable_class_helper.pxi.in

@@ -257,7 +257,7 @@ cdef class StringVector(Vector):
         self.data.n = 0
         self.data.m = _INIT_VEC_CAP
         self.data.data = <char **>malloc(self.data.m * sizeof(char *))
-        if not self.data.data:
+        if self.data.data is NULL:
             raise MemoryError()
 
     cdef resize(self):
@@ -270,7 +270,7 @@ cdef class StringVector(Vector):
 
         orig_data = self.data.data
         self.data.data = <char **>malloc(self.data.m * sizeof(char *))
-        if not self.data.data:
+        if self.data.data is NULL:
             raise MemoryError()
         for i in range(m):
             self.data.data[i] = orig_data[i]
@@ -975,6 +975,8 @@ cdef class StringHashTable(HashTable):
             const char **vecs
 
         vecs = <const char **>malloc(n * sizeof(char *))
+        if vecs is NULL:
+            raise MemoryError()
         for i in range(n):
             val = values[i]
             v = get_c_string(val)
@@ -1005,6 +1007,8 @@ cdef class StringHashTable(HashTable):
 
         # these by-definition *must* be strings
         vecs = <const char **>malloc(n * sizeof(char *))
+        if vecs is NULL:
+            raise MemoryError()
         for i in range(n):
             val = values[i]
 
@@ -1041,6 +1045,8 @@ cdef class StringHashTable(HashTable):
 
         # these by-definition *must* be strings
         vecs = <const char **>malloc(n * sizeof(char *))
+        if vecs is NULL:
+            raise MemoryError()
         for i in range(n):
             val = values[i]
 
@@ -1116,6 +1122,8 @@ cdef class StringHashTable(HashTable):
 
         # assign pointers and pre-filter out missing (if ignore_na)
         vecs = <const char **>malloc(n * sizeof(char *))
+        if vecs is NULL:
+            raise MemoryError()
         for i in range(n):
             val = values[i]
 

pandas/_libs/sas.pyx

@@ -49,7 +49,7 @@ cdef bytes buf_as_bytes(Buffer buf, size_t offset, size_t length):
 
 cdef Buffer buf_new(size_t length) except *:
     cdef uint8_t *data = <uint8_t *>calloc(length, sizeof(uint8_t))
-    if data == NULL:
+    if data is NULL:
         raise MemoryError(f"Failed to allocate {length} bytes")
     return Buffer(data, length)
 

pandas/_libs/tslibs/period.pyx

@@ -679,6 +679,8 @@ cdef char* c_strftime(npy_datetimestruct *dts, char *fmt):
     c_date.tm_isdst = -1
 
     result = <char*>malloc(result_len * sizeof(char))
+    if result is NULL:
+        raise MemoryError()
 
     strftime(result, result_len, fmt, &c_date)