Port to Py2/3. Test passes all but 2 in Py3 and all but 1 in Py2.

pyfribidi.c

@@ -21,17 +21,26 @@
 #include <Python.h>
 #include <fribidi.h>
 
+#if PY_MAJOR_VERSION >= 3
+#define PYFRIBIDI_UNICODE_OBJECT PyObject
+#else
+#define PYFRIBIDI_UNICODE_OBJECT PyUnicodeObject
+#endif
 
 static PyObject *
-unicode_log2vis (PyUnicodeObject* string,
+unicode_log2vis (PYFRIBIDI_UNICODE_OBJECT * string,
                  FriBidiParType base_direction, int clean, int reordernsm)
 {
     int i;
+#if PY_MAJOR_VERSION >= 3
+    Py_ssize_t length = PyUnicode_GetLength(string);
+#else
     int length = string->length;
+#endif
     FriBidiChar *logical = NULL; /* input fribidi unicode buffer */
     FriBidiChar *visual = NULL;      /* output fribidi unicode buffer */
-    FriBidiStrIndex new_len = 0; /* length of the UTF-8 buffer */
-    PyUnicodeObject *result = NULL;
+    // FriBidiStrIndex new_len = 0; /* length of the UTF-8 buffer */
+    PYFRIBIDI_UNICODE_OBJECT *result = NULL;
 
     /* Allocate fribidi unicode buffers
        TODO - Don't copy strings if sizeof(FriBidiChar) == sizeof(Py_UNICODE)
@@ -49,9 +58,14 @@ unicode_log2vis (PyUnicodeObject* string,
         goto cleanup;
     }
 
+#if PY_MAJOR_VERSION >= 3
+    int number_of_copied_wchar = PyUnicode_AsWideChar(string, logical, length);
+    //logical[number_of_copied_wchar] = (wchar_t) NULL;
+#else
     for (i=0; i<length; ++i) {
         logical[i] = string->str[i];
     }
+#endif
 
     /* Convert to unicode and order visually */
     fribidi_set_reorder_nsm(reordernsm);
@@ -69,14 +83,21 @@ unicode_log2vis (PyUnicodeObject* string,
         length = fribidi_remove_bidi_marks (visual, length, NULL, NULL, NULL);
     }
 
-    result = (PyUnicodeObject*) PyUnicode_FromUnicode(NULL, length);
+#if PY_MAJOR_VERSION >= 3
+    result = PyUnicode_FromWideChar(visual, length);
+
+#else
+    result = (PYFRIBIDI_UNICODE_OBJECT *) PyUnicode_FromUnicode(NULL, length);
+#endif
     if (result == NULL) {
         goto cleanup;
     }
-
+#if PY_MAJOR_VERSION >= 3
+#else
     for (i=0; i<length; ++i) {
         result->str[i] = visual[i];
     }
+#endif
 
   cleanup:
     /* Delete unicode buffers */
@@ -89,7 +110,7 @@ unicode_log2vis (PyUnicodeObject* string,
 static PyObject *
 _pyfribidi_log2vis (PyObject * self, PyObject * args, PyObject * kw)
 {
-    PyUnicodeObject *logical = NULL;	/* input unicode or string object */
+    PYFRIBIDI_UNICODE_OBJECT *logical = NULL;	/* input unicode or string object */
     FriBidiParType base = FRIBIDI_TYPE_RTL;	/* optional direction */
     int clean = 0; /* optional flag to clean the string */
     int reordernsm = 1; /* optional flag to allow reordering of non spacing marks*/
@@ -101,7 +122,6 @@ _pyfribidi_log2vis (PyObject * self, PyObject * args, PyObject * kw)
                                       &logical, &base, &clean, &reordernsm)) {
         return NULL;
     }
-
     /* Validate base */
 
     if (!(base == FRIBIDI_TYPE_RTL
@@ -112,7 +132,9 @@ _pyfribidi_log2vis (PyObject * self, PyObject * args, PyObject * kw)
                              base);
     }
 
-    return unicode_log2vis (logical, base, clean, reordernsm);
+    PyObject * return_pyobject = unicode_log2vis (logical, base, clean, reordernsm);
+
+    return return_pyobject;
 }
 
 
@@ -121,12 +143,40 @@ static PyMethodDef PyfribidiMethods[] = {
 	{NULL, NULL, 0, NULL}
 };
 
+#if PY_MAJOR_VERSION >= 3
+#define INITERROR return NULL
+
+static struct PyModuleDef pyfribidi_moduledef = {
+  PyModuleDef_HEAD_INIT, /* m_base */
+  "pyfribidi", /* m_name */
+  NULL, /* m_doc */
+  -1, /* m_size == -1 means cannot be reinitialized */
+  PyfribidiMethods, //
+  NULL, //  myextension_traverse,
+  NULL, //  myextension_clear,
+  NULL
+};
 
 PyMODINIT_FUNC
+PyInit__pyfribidi (void)
+#else
+#define INITERROR return
+
+void
 init_pyfribidi (void)
+#endif
 {
+#if PY_MAJOR_VERSION >= 3
+        PyObject *module = PyModule_Create (&pyfribidi_moduledef); // "_pyfribidi", PyfribidiMethods);
+#else
         PyObject *module = Py_InitModule ("_pyfribidi", PyfribidiMethods);
+#endif
+	if (module == NULL) INITERROR;
+
 	PyModule_AddIntConstant (module, "RTL", (long) FRIBIDI_TYPE_RTL);
 	PyModule_AddIntConstant (module, "LTR", (long) FRIBIDI_TYPE_LTR);
 	PyModule_AddIntConstant (module, "ON", (long) FRIBIDI_TYPE_ON);
+#if PY_MAJOR_VERSION >= 3
+	return module;
+#endif
 }

pyfribidi.py

@@ -5,6 +5,7 @@
 logical order, but the conversion may be wrong in certain cases.
 """
 
+from builtins import str
 __version__ = "0.11.0"
 from _pyfribidi import LTR, ON, RTL, log2vis as _log2vis
 
@@ -25,8 +26,8 @@ def log2vis(logical, base_direction=RTL, encoding="utf-8", clean=False, reordern
     - encoding: optional string encoding (ignored for unicode input)
     """
 
-    if not isinstance(logical, unicode):
-        logical = unicode(logical, encoding)
+    if not isinstance(logical, str):
+        logical = str(logical, encoding)
     else:
         encoding = None
     res = _log2vis(logical, base_direction=base_direction, clean=clean, reordernsm=reordernsm)

requirements-27.txt

@@ -0,0 +1,2 @@
+future==0.16.0
+six==1.10.0

requirements-36.txt

@@ -0,0 +1,2 @@
+future==0.16.0
+six==1.10.0

setup.py

@@ -1,12 +1,13 @@
 #!/usr/bin/env python
 
+from past.builtins import execfile
+import os
 from distutils.core import setup
 from distutils.extension import Extension
 from distutils.command import build_ext
-from my_build_ext import build_ext as _link_build_ext
 from distutils.util import strtobool
 
-import os
+from my_build_ext import build_ext as _link_build_ext
 
 EMBED = strtobool(os.environ.get("EMBED", "1"))
 

test_pyfribidi.py

@@ -8,6 +8,7 @@
 
 import sys
 import unittest
+
 import pyfribidi
 from pyfribidi import RTL, LTR, ON
 
@@ -21,7 +22,7 @@ class InputTests(unittest.TestCase):
     def testRequireInputString(self):
         """ input: require input string or unicode """
         self.assertRaises(TypeError, pyfribidi.log2vis)
-        
+
     def testInvalidInputString(self):
         """ input: raise TypeError for non string or unicode input """
         self.assertRaises(TypeError, pyfribidi.log2vis, 1)
@@ -40,8 +41,8 @@ def testInvalidEncodedString(self):
         """ input: raise UnicodeError for invalid encoded string """
         self.assertRaises(UnicodeError, pyfribidi.log2vis, "שלום",
                           encoding='iso8859-8')
-        
-        
+
+
 class UnicodeTests(unittest.TestCase):
 
     def testEmpty(self):
@@ -176,10 +177,15 @@ def testUTF16NaturalRTL(self):
                                            encoding=charset),
                          u"hello - םולש".encode(charset))
 
+
 class Crasher(unittest.TestCase):
     def test_glibc_free_invalid_next_size(self):
-        # *** glibc detected *** /home/ralf/py27/bin/python2: free(): invalid next size (fast): 0x00000000011cff00 ***
-        pyfribidi.log2vis('\xf0\x90\x8e\xa2\xf0\x90\x8e\xaf\xf0\x90\x8e\xb4\xf0\x90\x8e\xa1\xf0\x90\x8f\x83')
+        # *** glibc detected *** /home/ralf/py27/bin/python2: free(): invalid
+        # next size (fast): 0x00000000011cff00 ***
+        pyfribidi.log2vis(
+            '\xf0\x90\x8e\xa2\xf0\x90\x8e\xaf\xf0\x90\x8e\xb4\xf0\x90\x8e\xa1'
+            '\xf0\x90\x8f\x83')
+
 
 if __name__ == '__main__':
     suite = unittest.defaultTestLoader.loadTestsFromName(__name__)

time_pyfribidi.py

@@ -3,7 +3,9 @@
 """ 
 time pyfribidi with all hebrew encodings.
 """
+from __future__ import print_function
 
+from builtins import range
 import syspath
 import timeit
 import pyfribidi
@@ -30,24 +32,24 @@ def timeEncoding(encoding, tests):
     timer = timeit.Timer(code, setup)
     seconds = timer.timeit(number=tests)
     microseconds = 1000000 * seconds / tests
-    print "%12s: %.8f seconds (%.2f usec/pass)" % (encoding, seconds,
-                                                   microseconds)
+    print("%12s: %.8f seconds (%.2f usec/pass)" % (encoding, seconds,
+                                                   microseconds))
 
 
 # warm up caches
-for i in xrange(100000):
+for i in range(100000):
     pyfribidi.log2vis(u'Some text to warm up the caches')
 
 lines = 50 # typical screen of text
-print
-print 'time to rerorder %s lines:' % lines
-print
+print()
+print('time to rerorder %s lines:' % lines)
+print()
 for encoding in hebrew_encodings:
     timeEncoding(encoding, lines)
 
 lines = 100000 
-print
-print 'time to rerorder %s lines:' % lines
-print
+print()
+print('time to rerorder %s lines:' % lines)
+print()
 for encoding in hebrew_encodings:
     timeEncoding(encoding, lines)