Merge pull request #25 from dw/py3-only

Move to Python 3 only

Author: bbayles

.travis.yml

@@ -2,12 +2,10 @@ dist: xenial
 language: "python"
 
 python:
-    - "2.7"
     - "3.4"
     - "3.5"
     - "3.6"
     - "3.7"
-    - "pypy2.7-6.0"
     - "pypy3.5-6.0"
 
 install:
@@ -16,6 +14,7 @@ install:
 
 script:
     - "coverage run --include='cdblib/*.py' setup.py test"
+    - "ENABLE_DJB_HASH_CEXT=1 coverage run --include='cdblib/*.py' setup.py test"
     - "flake8 ."
 
 notifications:

README.rst

@@ -8,13 +8,15 @@ python-pure-cdb
     :target: https://python-pure-cdb.readthedocs.io/en/latest/?badge=latest
 
 The `python-pure-cdb` package (`pure-cdb <https://pypi.org/project/pure-cdb/>`_ on PyPI)
-is a Python library for working with D.J. Bernstein's "constant datbases."
-It supports both Python 2.7 and Python 3.
+is a Python library for working with D.J. Bernstein's "constant databases."
 
 In addition to being able to read and write the database files produced by
-other `cdb` tools, `python-pure-cdb` can produce and consume "64-bit"
+other `cdb` tools, this package can produce and consume "64-bit"
 constant databases that don't have the usual 4 GiB restriction.
 
+This package works with Python 3.4 and above.
+For a version that works with Python 2, see `this older release <https://github.com/dw/python-pure-cdb/releases/tag/v2.2.0>`_.
+
 For more information on constant databases, see `djb's page <https://cr.yp.to/cdb.html>`_
 and `Wikipedia <https://en.wikipedia.org/wiki/Cdb_(software)>`_.
 

cdblib/__init__.py

@@ -1,5 +1,3 @@
-from __future__ import unicode_literals
-
 from .djb_hash import djb_hash
 from .cdblib import Reader, Reader64, Writer, Writer64
 

cdblib/_djb_hash.c

@@ -1,15 +1,9 @@
 #define PY_SSIZE_T_CLEAN
 #include "Python.h"
 
-#if PY_MAJOR_VERSION >= 3
-#   define MOD_RETURN(mod) return mod;
-#   define MODINIT_NAME PyInit__djb_hash
-#   define BUFFERLIKE_FMT "y#"
-#else
-#   define MOD_RETURN(mod) return;
-#   define MODINIT_NAME init_djb_hash
-#   define BUFFERLIKE_FMT "t#"
-#endif
+#define MOD_RETURN(mod) return mod;
+#define MODINIT_NAME PyInit__djb_hash
+#define BUFFERLIKE_FMT "y#"
 
 /* Return a Python long instance containing the value of the DJB hash function
  * for the given string or array. */
@@ -36,7 +30,6 @@ static /*const*/ PyMethodDef module_methods[] = {
     {NULL, NULL, 0, NULL}
 };
 
-#if PY_MAJOR_VERSION >= 3
 static struct PyModuleDef moduledef = {
     PyModuleDef_HEAD_INIT,
     "_djb_hash",
@@ -48,16 +41,11 @@ static struct PyModuleDef moduledef = {
     NULL,
     NULL
 };
-#endif
 
 PyMODINIT_FUNC
 MODINIT_NAME(void)
 {
-#if PY_MAJOR_VERSION >= 3
     PyObject *mod = PyModule_Create(&moduledef);
-#else
-    PyObject *mod = Py_InitModule3("_djb_hash", module_methods, "");
-#endif
 
     MOD_RETURN(mod);
 }

cdblib/cdbdump.py

@@ -1,23 +1,19 @@
-from __future__ import print_function
-
 import argparse
 import sys
 
-import six
-
 import cdblib
 
 
 def cdbdump(parsed_args, **kwargs):
     # Read binary data from stdin by default
     stdin = kwargs.get('stdin')
     if stdin is None:
-        stdin = sys.stdin if six.PY2 else sys.stdin.buffer
+        stdin = sys.stdin.buffer
 
     # Print text data to stdout by default
     stdout = kwargs.get('stdout')
     if stdout is None:
-        stdout = sys.stdout if six.PY2 else sys.stdout.buffer
+        stdout = sys.stdout.buffer
 
     # Consume stdin and parse the cdb file
     reader_cls = cdblib.Reader64 if parsed_args['64'] else cdblib.Reader

cdblib/cdblib.py

@@ -5,14 +5,9 @@
     http://cr.yp.to/cdb.html
 
 '''
-from __future__ import unicode_literals
-
 from struct import Struct
 from itertools import chain
 
-import six
-from six.moves import range
-
 from .djb_hash import djb_hash
 
 # Structs for 32-bit databases
@@ -24,9 +19,10 @@
 write_2_le8 = Struct('<QQ').pack
 
 # Encoders for keys
-DEFAULT_ENCODERS = {six.text_type: lambda x: x.encode('utf-8')}
-for t in six.integer_types:
-    DEFAULT_ENCODERS[t] = lambda x: six.text_type(x).encode('utf-8')
+DEFAULT_ENCODERS = {
+    str: lambda x: x.encode('utf-8'),
+    int: lambda x: str(x).encode('utf-8'),
+}
 
 
 class _CDBBase(object):
@@ -41,7 +37,7 @@ def __init__(self, hashfn=djb_hash, strict=False, encoders=None):
             self.encoders.update(encoders)
 
     def hash_key(self, key):
-        if not isinstance(key, six.binary_type):
+        if not isinstance(key, bytes):
             try:
                 encoded_key = self.encoders[type(key)](key)
             except KeyError as e:
@@ -58,8 +54,8 @@ def hash_key_strict(self, key):
         try:
             h = self.hashfn(key)
         except TypeError as e:
-            msg = 'key must be of type {}'
-            e.args = (msg.format(six.binary_type.__name__),)
+            msg = 'key must be of type bytes'
+            e.args = (msg,)
             raise
 
         # Truncate to 32 bits and remove sign.
@@ -232,9 +228,8 @@ def __exit__(self, exc_type, exc_value, traceback):
     def put(self, key, value=b''):
         '''Write a string key/value pair to the output file.'''
         # Ensure that the value is binary
-        if not isinstance(value, six.binary_type):
-            msg = 'value must be of type {}'
-            raise TypeError(msg.format(six.binary_type.__name__))
+        if not isinstance(value, bytes):
+            raise TypeError('value must be of type bytes')
 
         # Computing the hash for the key also ensures that it's binary
         key, h = self.hash_key(key)
@@ -265,12 +260,12 @@ def putints(self, key, values):
     def putstring(self, key, value, encoding='utf-8'):
         '''Write a unicode string associated with the given key to the output
         file after encoding it as UTF-8 or the given encoding.'''
-        self.put(key, six.text_type.encode(value, encoding))
+        self.put(key, value.encode(encoding))
 
     def putstrings(self, key, values, encoding='utf-8'):
         '''Write zero or more unicode strings to the output file. Equivalent to
         calling putstring() in a loop.'''
-        self.puts(key, (six.text_type.encode(v, encoding) for v in values))
+        self.puts(key, (v.encode(encoding) for v in values))
 
     def finalize(self):
         '''Write the final hash tables to the output file, and write out its

cdblib/cdbmake.py

@@ -1,12 +1,7 @@
-from __future__ import print_function, unicode_literals
-
 import argparse
-import io
 import os
 import sys
 
-import six
-
 import cdblib
 
 
@@ -15,7 +10,7 @@ def __init__(self, parsed_args, **kwargs):
         # Read binary data from stdin and write errors to stderr (by default)
         self.stdin = kwargs.get('stdin')
         if self.stdin is None:
-            self.stdin = sys.stdin if six.PY2 else sys.stdin.buffer
+            self.stdin = sys.stdin.buffer
 
         self.stderr = kwargs.get('stderr', sys.stderr)
 
@@ -93,7 +88,7 @@ def get_items(self):
             yield key, data
 
     def run(self):
-        with io.open(self.cdb_temp_path, 'wb') as tmpfile:
+        with open(self.cdb_temp_path, 'wb') as tmpfile:
             with self.writer_cls(tmpfile) as writer:
                 for key, data in self.get_items():
                     writer.put(key, data)

cdblib/djb_hash.py

@@ -1,26 +1,13 @@
-from __future__ import unicode_literals
-
-import six
-
 # The cdb hash function is defined at: http://cr.yp.to/cdb/cdb.txt
-# We use separate implementations for Python 2 and 3, since they treat
-# iterating over byte strings differently.
-if six.PY2:
-    def djb_hash(s):
-        '''Return the value of DJB's hash function for byte string *s*'''
-        h = 5381
-        for c in s:
-            h = (((h << 5) + h) ^ ord(c)) & 0xffffffff
-        return h
-else:
-    def djb_hash(s):  # noqa
-        '''Return the value of DJB's hash function for byte string *s*'''
-        h = 5381
-        for c in s:
-            h = (((h << 5) + h) ^ c) & 0xffffffff
-        return h
+def djb_hash(s):  # noqa
+    '''Return the value of DJB's hash function for byte string *s*'''
+    h = 5381
+    for c in s:
+        h = (((h << 5) + h) ^ c) & 0xffffffff
+    return h
+
 
-# If the C Extensions is available (Python 2 only), use it
+# If the C Extension is available, use it
 try:
     from ._djb_hash import djb_hash  # noqa
 except ImportError:

docs/conf.py

@@ -26,7 +26,7 @@
 # The short X.Y version
 version = ''
 # The full version, including alpha/beta/rc tags
-release = '2.2.0'
+release = '3.0.0'
 
 
 # -- General configuration ---------------------------------------------------

docs/library.rst

@@ -119,9 +119,7 @@ Encoding and strict mode
 ^^^^^^^^^^^^^^^^^^^^^^^^
 
 Database keys are stored as `bytes` objects. By default, `Reader` instances
-will attempt to convert text keys (`str` on Python 3, `unicode` on Python 2)
-and integer keys (`int` on Python 3, `int` and `long` on Python 2)
-automatically.
+will attempt to convert `str` keys and `int` keys automatically.
 
     >>> reader.get(b'1')  # Binary key
     b'value_for_1'