Remove Python 2 compatibility shims (#34)

* Remove Python 2 compatibility shims

Python 2 compatibility was removed in #31 so these should not be useful / necessary anymore.

* Remove dead `sys` import

The CLI mode uses `sys` internally but has its own import.

* Remove `to_file_bytes`

It does almost nothing at this point and is only used by one caller, so can be inlined.

* Minor fixes for Python 3 compat

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

* Fix nested next method

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

* Validate applying 30MB patch file

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

* Update CI to fix hanging Linux build

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

* Use specific patch version for python due python-setup rule

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

* Use ubuntu 22.04 to have python 3.7 available

Signed-off-by: Uilian Ries <uilianr@jfrog.com>

---------

Signed-off-by: Uilian Ries <uilianr@jfrog.com>
Co-authored-by: Uilian Ries <uilianr@jfrog.com>

Author: xmo-odoo

.github/workflows/workflow.yml

@@ -21,15 +21,15 @@ on:
 jobs:
   linux-validate:
     name: Validate on Linux - Python ${{ matrix.python }}
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-22.04
     strategy:
       matrix:
-        python: [ '3.6', '3.8', '3.12' ]
+        python: [ '3.7', '3.8', '3.12' ]
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
 
       - name: Setup python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: ${{ matrix.python }}
           architecture: x64
@@ -42,12 +42,12 @@ jobs:
     runs-on: windows-latest
     strategy:
       matrix:
-        python: [ '3.6', '3.8', '3.12' ]
+        python: [ '3.7', '3.8', '3.12' ]
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
 
       - name: Setup python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: ${{ matrix.python }}
           architecture: x64

example/example.py

@@ -8,7 +8,7 @@ def __init__(self):
 
     def emit(self, record):
         logstr = self.format(record)
-        print logstr
+        print(logstr)
 
 patchlog = logging.getLogger("patch")
 patchlog.handlers = []

patch_ng.py

@@ -28,56 +28,23 @@
     ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
     SOFTWARE.
 """
-from __future__ import print_function
-
 __author__ = "Conan.io <info@conan.io>"
 __version__ = "1.19.0-dev"
 __license__ = "MIT"
 __url__ = "https://github.com/conan-io/python-patch"
 
+import codecs
 import copy
+import io
 import logging
-import re
-import tempfile
-import codecs
-
-# cStringIO doesn't support unicode in 2.5
-try:
-  from StringIO import StringIO
-except ImportError:
-  from io import BytesIO as StringIO # python 3
-try:
-  import urllib2 as urllib_request
-except ImportError:
-  import urllib.request as urllib_request
-
-from os.path import exists, isfile, abspath
 import os
 import posixpath
+import re
 import shutil
-import sys
 import stat
-
-
-PY3K = sys.version_info >= (3, 0)
-
-# PEP 3114
-if not PY3K:
-  compat_next = lambda gen: gen.next()
-else:
-  compat_next = lambda gen: gen.__next__()
-
-def tostr(b):
-  """ Python 3 bytes encoder. Used to print filename in
-      diffstat output. Assumes that filenames are in utf-8.
-  """
-  if not PY3K:
-    return b
-
-  # [ ] figure out how to print non-utf-8 filenames without
-  #     information loss
-  return b.decode('utf-8')
-
+import tempfile
+import urllib.request
+from os.path import exists, isfile, abspath
 
 #------------------------------------------------
 # Logging is controlled by logger named after the
@@ -90,22 +57,10 @@ def tostr(b):
 warning = logger.warning
 error = logger.error
 
-class NullHandler(logging.Handler):
-  """ Copied from Python 2.7 to avoid getting
-      `No handlers could be found for logger "patch"`
-      http://bugs.python.org/issue16539
-  """
-  def handle(self, record):
-    pass
-  def emit(self, record):
-    pass
-  def createLock(self):
-    self.lock = None
-
 streamhandler = logging.StreamHandler()
 
 # initialize logger itself
-logger.addHandler(NullHandler())
+logger.addHandler(logging.NullHandler())
 
 debugmode = False
 
@@ -194,19 +149,18 @@ def fromfile(filename):
   """
   patchset = PatchSet()
   debug("reading %s" % filename)
-  fp = open(filename, "rb")
-  res = patchset.parse(fp)
-  fp.close()
-  if res == True:
-    return patchset
+  with open(filename, "rb") as fp:
+    res = patchset.parse(fp)
+    if res == True:
+      return patchset
   return False
 
 
 def fromstring(s):
   """ Parse text string and return PatchSet()
       object (or False if parsing fails)
   """
-  ps = PatchSet( StringIO(s) )
+  ps = PatchSet( io.BytesIO(s) )
   if ps.errors == 0:
     return ps
   return False
@@ -217,7 +171,7 @@ def fromurl(url):
       if an error occured. Note that this also
       can throw urlopen() exceptions.
   """
-  ps = PatchSet( urllib_request.urlopen(url) )
+  ps = PatchSet( urllib.request.urlopen(url) )
   if ps.errors == 0:
     return ps
   return False
@@ -261,15 +215,6 @@ def decode_text(text):
   return text.decode("utf-8", "ignore")  # Ignore not compatible characters
 
 
-def to_file_bytes(content):
-  if PY3K:
-    if not isinstance(content, bytes):
-      content = bytes(content, "utf-8")
-  elif isinstance(content, unicode):
-    content = content.encode("utf-8")
-  return content
-
-
 def load(path, binary=False):
   """ Loads a file content """
   with open(path, 'rb') as handle:
@@ -290,7 +235,9 @@ def save(path, content, only_if_modified=False):
   except Exception:
     pass
 
-  new_content = to_file_bytes(content)
+  new_content = content
+  if not isinstance(content, bytes):
+      new_content = bytes(content, "utf-8")
 
   if only_if_modified and os.path.exists(path):
     old_content = load(path, binary=True)
@@ -328,8 +275,7 @@ def __init__(self):
     self.type = None
 
   def __iter__(self):
-    for h in self.hunks:
-      yield h
+    return iter(self.hunks)
 
 
 class PatchSet(object):
@@ -359,8 +305,7 @@ def __len__(self):
     return len(self.items)
 
   def __iter__(self):
-    for i in self.items:
-      yield i
+    return iter(self.items)
 
   def parse(self, stream):
     """ parse unified diff
@@ -394,7 +339,7 @@ def next(self):
           return False
 
         try:
-          self._lineno, self._line = compat_next(super(wrapumerate, self))
+          self._lineno, self._line = super(wrapumerate, self).__next__()
         except StopIteration:
           self._exhausted = True
           self._line = False
@@ -902,7 +847,7 @@ def diffstat(self):
         #print(iratio, dratio, iwidth, dwidth, histwidth)
         hist = "+"*int(iwidth) + "-"*int(dwidth)
       # -- /calculating +- histogram --
-      output += (format % (tostr(names[i]), str(insert[i] + delete[i]), hist))
+      output += (format % (names[i].decode('utf-8'), str(insert[i] + delete[i]), hist))
 
     output += (" %d files changed, %d insertions(+), %d deletions(-), %+d bytes"
                % (len(names), sum(insert), sum(delete), delta))
@@ -1270,15 +1215,11 @@ def get_line():
 
 
   def write_hunks(self, srcname, tgtname, hunks):
-    src = open(srcname, "rb")
-    tgt = open(tgtname, "wb")
-
-    debug("processing target file %s" % tgtname)
-
-    tgt.writelines(self.patch_stream(src, hunks))
+    with open(srcname, "rb") as src, open(tgtname, "wb") as tgt:
+        debug("processing target file %s" % tgtname)
+    
+        tgt.writelines(self.patch_stream(src, hunks))
 
-    tgt.close()
-    src.close()
     # [ ] TODO: add test for permission copy
     shutil.copymode(srcname, tgtname)
     return True

tests/06nested/tests/app/EVENT_LOOP.py

@@ -21,8 +21,8 @@
 
 class EVENT_LOOP(unittest.TestCase):
     def t_scheduled(self, interval, iterations, sleep_time=0):
-        print 'Test interval=%s, iterations=%s, sleep=%s' % (interval,
-            iterations, sleep_time)
+        print('Test interval=%s, iterations=%s, sleep=%s' % (interval,
+            iterations, sleep_time))
         warmup_iterations = iterations
 
         self.last_t = 0.
@@ -76,6 +76,6 @@ def test_d01_50(self):
 
 if __name__ == '__main__':
     if pyglet.version != '1.2dev':
-        print 'Wrong version of pyglet imported; please check your PYTHONPATH'
+        print('Wrong version of pyglet imported; please check your PYTHONPATH')
     else:
         unittest.main()

tests/run_tests.py

@@ -36,15 +36,10 @@
 import re
 import shutil
 import unittest
-import copy
 import stat
 from os import listdir, chmod
 from os.path import abspath, dirname, exists, join, isdir, isfile
 from tempfile import mkdtemp
-try:
-  getcwdu = os.getcwdu
-except AttributeError:
-  getcwdu = os.getcwd  # python 3, where getcwd always returns a unicode object
 
 verbose = False
 if "-v" in sys.argv or "--verbose" in sys.argv:
@@ -150,7 +145,7 @@ def _run_test(self, testname):
       # 3.
       # test utility as a whole
       patch_tool = join(dirname(TESTS), "patch_ng.py")
-      save_cwd = getcwdu()
+      save_cwd = os.getcwd()
       os.chdir(tmpdir)
       extra = "-f" if "10fuzzy" in testname else ""
       if verbose:
@@ -204,7 +199,7 @@ def create_closure():
 
 class TestCheckPatched(unittest.TestCase):
     def setUp(self):
-        self.save_cwd = getcwdu()
+        self.save_cwd = os.getcwd()
         os.chdir(TESTS)
 
     def tearDown(self):
@@ -355,7 +350,7 @@ def test(self):
 
 class TestPatchApply(unittest.TestCase):
     def setUp(self):
-        self.save_cwd = getcwdu()
+        self.save_cwd = os.getcwd()
         self.tmpdir = mkdtemp(prefix=self.__class__.__name__)
         os.chdir(self.tmpdir)
 
@@ -457,6 +452,41 @@ def test_unlink_backup_windows(self):
         self.assertTrue(os.stat(some_file).st_mode, stat.S_IRUSR | stat.S_IRGRP | stat.S_IROTH)
 
 
+class TestHugePatchFile(unittest.TestCase):
+
+    def setUp(self):
+        self.save_cwd = os.getcwd()
+        self.tmpdir = mkdtemp(prefix=self.__class__.__name__)
+        os.chdir(self.tmpdir)
+        self.huge_patchfile = self._create_huge_patchfile(self.tmpdir)
+
+    def _create_huge_patchfile(self, tmpdir):
+        """ Create a patch file with ~30MB of data
+        """
+        hugefile = join(tmpdir, 'hugefile')
+        with open(hugefile, 'wb') as f:
+            for i in range(2500000):
+                f.write(b'Line %d\n' % i)
+        huge_patchfile = join(tmpdir, 'huge.patch')
+        with open(huge_patchfile, 'wb') as f:
+            f.write(b'--- a/hugefile\n')
+            f.write(b'+++ b/hugefile\n')
+            f.write(b'@@ -1,2500000 +1,2500000 @@\n')
+            for i in range(2500000):
+                f.write(b' Line %d\n' % i)
+        return huge_patchfile
+
+    def tearDown(self):
+        os.chdir(self.save_cwd)
+        remove_tree_force(self.tmpdir)
+
+    def test_apply_huge_patch(self):
+        """ Test that a huge patch file can be applied without issues
+        """
+        pto = patch_ng.fromfile(self.huge_patchfile)
+        self.assertTrue(pto.apply(root=self.tmpdir))
+
+
 class TestHelpers(unittest.TestCase):
     # unittest setting
     longMessage = True