Fix "std::string::npos" in 'libcpp.string' and add missing C++ string methods (GH-4276)

Closes https://github.com/cython/cython/issues/4268

Cython/Includes/libcpp/string.pxd

@@ -2,16 +2,15 @@
 # deprecated cimport for backwards compatibility:
 from libc.string cimport const_char
 
+cdef extern from "<string>" namespace "std::string" nogil:
+    const size_t npos
 
 cdef extern from "<string>" namespace "std" nogil:
-
-    size_t npos = -1
-
     cdef cppclass string:
         cppclass iterator:
             iterator()
             char& operator*()
-            iterator(iterator &)
+            iterator(iterator&)
             iterator operator++()
             iterator operator--()
             bint operator==(iterator)
@@ -34,12 +33,12 @@ cdef extern from "<string>" namespace "std" nogil:
             pass
 
         string() except +
-        string(const char *) except +
-        string(const char *, size_t) except +
-        string(const string&) except +
-        # as a string formed by a repetition of character c, n times.
-        string(size_t, char) except +
-        # from a pair of iterators
+        string(const string& s) except +
+        string(const string& s, size_t pos) except +
+        string(const string& s, size_t pos, size_t len) except +
+        string(const char* s) except +
+        string(const char* s, size_t n) except +
+        string(size_t n, char c) except +
         string(iterator first, iterator last) except +
 
         iterator begin()
@@ -56,111 +55,122 @@ cdef extern from "<string>" namespace "std" nogil:
         size_t size()
         size_t max_size()
         size_t length()
-        void resize(size_t)
-        void resize(size_t, char c)
+        void resize(size_t) except +
+        void resize(size_t, char) except +
+        void shrink_to_fit() except +
         size_t capacity()
-        void reserve(size_t)
+        void reserve(size_t) except +
         void clear()
         bint empty()
-        iterator erase(iterator position)
-        iterator erase(const_iterator position)
         iterator erase(iterator first, iterator last)
+        iterator erase(iterator p)
         iterator erase(const_iterator first, const_iterator last)
-
-        char& at(size_t) except +
-        char& operator[](size_t)
-        char& front()  # C++11
-        char& back()   # C++11
-        int compare(const string&)
-
-        string& append(const string&) except +
-        string& append(const string&, size_t, size_t) except +
-        string& append(const char *) except +
-        string& append(const char *, size_t) except +
-        string& append(size_t, char) except +
+        iterator erase(const_iterator p)
+        string& erase(size_t pos, size_t len) except +
+        string& erase(size_t pos) except +
+        string& erase() except +
+
+        char& at(size_t pos) except +
+        char& operator[](size_t pos)
+        char& front()
+        char& back()
+        int compare(const string& s)
+        int compare(size_t pos, size_t len, const string& s) except +
+        int compare(size_t pos, size_t len, const string& s, size_t subpos, size_t sublen) except +
+        int compare(const char* s) except +
+        int compare(size_t pos, size_t len, const char* s) except +
+        int compare(size_t pos, size_t len, const char* s , size_t n) except +
+
+        string& append(const string& s) except +
+        string& append(const string& s, size_t subpos, size_t sublen) except +
+        string& append(const char* s) except +
+        string& append(const char* s, size_t n) except +
+        string& append(size_t n, char c) except +
 
         void push_back(char c) except +
-
-        string& assign (const string&)
-        string& assign (const string&, size_t, size_t)
-        string& assign (const char *, size_t)
-        string& assign (const char *)
-        string& assign (size_t n, char c)
-
-        string& insert(size_t, const string&) except +
-        string& insert(size_t, const string&, size_t, size_t) except +
-        string& insert(size_t, const char* s, size_t) except +
-
-
-        string& insert(size_t, const char* s) except +
-        string& insert(size_t, size_t, char c) except +
-
-        size_t copy(char *, size_t, size_t) except +
-
-        size_t find(const string&, size_t pos)
-        size_t find(const string&)
-        size_t find(const char*, size_t pos, size_t n)
-        size_t find(const char*, size_t pos)
-        size_t find(const char*)
+        void pop_back()
+
+        string& assign (const string& s) except +
+        string& assign (const string& s, size_t subpos, size_t sublen) except +
+        string& assign (const char* s, size_t n) except +
+        string& assign (const char* s) except +
+        string& assign (size_t n, char c) except +
+
+        string& insert(size_t pos, const string& s, size_t subpos, size_t sublen) except +
+        string& insert(size_t pos, const string& s) except +
+        string& insert(size_t pos, const char* s, size_t n) except +
+        string& insert(size_t pos, const char* s) except +
+        string& insert(size_t pos, size_t n, char c) except +
+        void insert(iterator p, size_t n, char c) except +
+        iterator insert(iterator p, char c) except +
+
+        size_t copy(char* s, size_t len, size_t pos) except +
+        size_t copy(char* s, size_t len) except +
+
+        size_t find(const string& s, size_t pos)
+        size_t find(const string& s)
+        size_t find(const char* s, size_t pos, size_t n)
+        size_t find(const char* s, size_t pos)
+        size_t find(const char* s)
         size_t find(char c, size_t pos)
         size_t find(char c)
 
         size_t rfind(const string&, size_t pos)
         size_t rfind(const string&)
         size_t rfind(const char* s, size_t pos, size_t n)
-        size_t rfind(const char*, size_t pos)
-        size_t rfind(const char*)
+        size_t rfind(const char* s, size_t pos)
+        size_t rfind(const char* s)
         size_t rfind(char c, size_t pos)
         size_t rfind(char c)
 
         size_t find_first_of(const string&, size_t pos)
         size_t find_first_of(const string&)
         size_t find_first_of(const char* s, size_t pos, size_t n)
-        size_t find_first_of(const char*, size_t pos)
-        size_t find_first_of(const char*)
+        size_t find_first_of(const char* s, size_t pos)
+        size_t find_first_of(const char* s)
         size_t find_first_of(char c, size_t pos)
         size_t find_first_of(char c)
 
-        size_t find_first_not_of(const string&, size_t pos)
-        size_t find_first_not_of(const string&)
-        size_t find_first_not_of(const char* s, size_t, size_t)
-        size_t find_first_not_of(const char*, size_t pos)
+        size_t find_first_not_of(const string& s, size_t pos)
+        size_t find_first_not_of(const string& s)
+        size_t find_first_not_of(const char* s, size_t pos, size_t n)
+        size_t find_first_not_of(const char* s, size_t pos)
         size_t find_first_not_of(const char*)
         size_t find_first_not_of(char c, size_t pos)
         size_t find_first_not_of(char c)
 
-        size_t find_last_of(const string&, size_t pos)
-        size_t find_last_of(const string&)
+        size_t find_last_of(const string& s, size_t pos)
+        size_t find_last_of(const string& s)
         size_t find_last_of(const char* s, size_t pos, size_t n)
-        size_t find_last_of(const char*, size_t pos)
-        size_t find_last_of(const char*)
+        size_t find_last_of(const char* s, size_t pos)
+        size_t find_last_of(const char* s)
         size_t find_last_of(char c, size_t pos)
         size_t find_last_of(char c)
 
-        size_t find_last_not_of(const string&, size_t pos)
-        size_t find_last_not_of(const string&)
+        size_t find_last_not_of(const string& s, size_t pos)
+        size_t find_last_not_of(const string& s)
         size_t find_last_not_of(const char* s, size_t pos, size_t n)
-        size_t find_last_not_of(const char*, size_t pos)
-        size_t find_last_not_of(const char*)
+        size_t find_last_not_of(const char* s, size_t pos)
+        size_t find_last_not_of(const char* s)
         size_t find_last_not_of(char c, size_t pos)
         size_t find_last_not_of(char c)
 
-        string substr(size_t, size_t) except +
+        string substr(size_t pos, size_t len) except +
+        string substr(size_t pos) except +
         string substr()
-        string substr(size_t) except +
 
         #string& operator= (const string&)
         #string& operator= (const char*)
         #string& operator= (char)
 
-        string operator+ (const string& rhs) except +
-        string operator+ (const char* rhs) except +
+        string operator+ (const string&) except +
+        string operator+ (const char*) except +
 
         bint operator==(const string&)
         bint operator==(const char*)
 
-        bint operator!= (const string& rhs )
-        bint operator!= (const char* )
+        bint operator!= (const string&)
+        bint operator!= (const char*)
 
         bint operator< (const string&)
         bint operator< (const char*)
@@ -175,12 +185,38 @@ cdef extern from "<string>" namespace "std" nogil:
         bint operator>= (const char*)
 
 
-    string to_string(int val)
-    string to_string(long val)
-    string to_string (long long val)
-    string to_string (unsigned val)
-    string to_string (unsigned long val)
-    string to_string (unsigned long long val)
-    string to_string (float val)
-    string to_string (double val)
-    string to_string (long double val)
+    string to_string(int val) except +
+    string to_string(long val) except +
+    string to_string(long long val) except +
+    string to_string(unsigned val) except +
+    string to_string(size_t val) except +
+    string to_string(ssize_t val) except +
+    string to_string(unsigned long val) except +
+    string to_string(unsigned long long val) except +
+    string to_string(float val) except +
+    string to_string(double val) except +
+    string to_string(long double val) except +
+
+    int stoi(const string& s, size_t* idx, int base) except +
+    int stoi(const string& s, size_t* idx) except +
+    int stoi(const string& s) except +
+    long stol(const string& s, size_t* idx, int base) except +
+    long stol(const string& s, size_t* idx) except +
+    long stol(const string& s) except +
+    long long stoll(const string& s, size_t* idx, int base) except +
+    long long stoll(const string& s, size_t* idx) except +
+    long long stoll(const string& s) except +
+
+    unsigned long stoul(const string& s, size_t* idx, int base) except +
+    unsigned long stoul(const string& s, size_t* idx) except +
+    unsigned long stoul(const string& s) except +
+    unsigned long long stoull(const string& s, size_t* idx, int base) except +
+    unsigned long long stoull(const string& s, size_t* idx) except +
+    unsigned long long stoull(const string& s) except +
+
+    float stof(const string& s, size_t* idx) except +
+    float stof(const string& s) except +
+    double stod(const string& s, size_t* idx) except +
+    double stod(const string& s) except +
+    long double stold(const string& s, size_t* idx) except +
+    long double stold(const string& s) except +

tests/run/cpp_stl_string.pyx

@@ -3,7 +3,7 @@
 
 cimport cython
 
-from libcpp.string cimport string, to_string
+from libcpp.string cimport string, npos, to_string, stoi, stof
 
 b_asdf = b'asdf'
 b_asdg = b'asdg'
@@ -102,6 +102,15 @@ def test_push_back(char *a):
     s.push_back(<char>ord('s'))
     return s.c_str()
 
+def test_pop_back(char *a):
+    """
+    >>> test_pop_back(b'abc') == b'ab' or test_pop_back(b'abc')
+    True
+    """
+    cdef string s = string(a)
+    s.pop_back()
+    return s
+
 def test_insert(char *a, char *b, int i):
     """
     >>> test_insert('AAAA'.encode('ASCII'), 'BBBB'.encode('ASCII'), 2) == 'AABBBBAA'.encode('ASCII')
@@ -133,6 +142,17 @@ def test_find(char *a, char *b):
     cdef size_t i = s.find(t)
     return i
 
+def test_npos(char *a, char *b):
+    """
+    >>> test_npos(b'abc', b'x')
+    True
+    >>> test_npos(b'abc', b'a')
+    False
+    """
+    cdef string s = string(a)
+    cdef string st = string(b)
+    return s.find(st) == npos
+
 def test_clear():
     """
     >>> test_clear() == ''.encode('ASCII')
@@ -142,6 +162,18 @@ def test_clear():
     s.clear()
     return s.c_str()
 
+def test_erase(char *a, size_t pos=0, size_t count=npos):
+    """
+    >>> test_erase(b'abc') == b'' or test_erase(b'abc')
+    True
+    >>> test_erase(b'abc', 1) == b'a' or test_erase(b'abc', 1)
+    True
+    >>> test_erase(b'abc', 1, 1) == b'ac' or test_erase(b'abc', 1, 1)
+    True
+    """
+    cdef string s = string(a)
+    return s.erase(pos, count)
+
 def test_assign(char *a):
     """
     >>> test_assign(b_asdf) == 'ggg'.encode('ASCII')
@@ -345,18 +377,34 @@ def test_iteration(string s):
     """
     return [c for c in s]
 
-
 def test_to_string(x):
     """
     >>> print(test_to_string(5))
-    si=5 sl=5
+    si=5 sl=5 ss=5 sss=5
     >>> print(test_to_string(-5))
-    si=-5 sl=-5
+    si=-5 sl=-5 ss=5 sss=-5
     """
     si = to_string(<int>x).decode('ascii')
     sl = to_string(<long>x).decode('ascii')
-    return f"si={si} sl={sl}"
+    ss = to_string(<size_t>abs(x)).decode('ascii')
+    sss = to_string(<ssize_t>x).decode('ascii')
+    return f"si={si} sl={sl} ss={ss} sss={sss}"
 
+def test_stoi(char *a):
+    """
+    >>> test_stoi(b'5')
+    5
+    """
+    cdef string s = string(a)
+    return stoi(s)
+
+def test_stof(char *a):
+    """
+    >>> test_stof(b'5.5')
+    5.5
+    """
+    cdef string s = string(a)
+    return stof(s)
 
 _WARNINGS = """
 21:31: Cannot pass Python object as C++ data structure reference (string &), will pass by copy.