Merge branch 'master' of https://github.com/python-pillow/Pillow into jpegdpi

CHANGES.rst

@@ -4,12 +4,63 @@ Changelog (Pillow)
 3.4.0 (unreleased)
 ------------------
 
+- Resampling lookups, trailing empty coefficients, precision #2008
+  [homm]
+
+- Add (un)packing between RGBA and BGRa #2057
+  [arjennienhuis]
+
+- Added return for J2k (and fpx) Load to return a pixel access object #2061
+  [wiredfool]
+
+- Skip failing numpy tests on Pypy <= 5.3.1 #2090
+  [arjennienhuis]
+  
+- Show warning when trying to save RGBA image as JPEG #2010
+  [homm]
+
+- Respect pixel centers during transform #2022
+  [homm]
+
+- TOC for supported file formats #2056
+  [polarize]
+
+- Fix conversion of bit images to numpy arrays Fixes #350, #2058
+  [matthew-brett]
+
+- Add ImageOps.scale to expand or contract a PIL image by a factor #2011
+  [vlmath]
+  
+- Flake8 fixes #2050
+  [hugovk]
+
+- Updated freetype to 2.6.5 on Appveyor builds #2035
+  [radarhere]
+  
+- PCX encoder fixes #2023, pr #2041
+  [homm]
+  
+- Docs: Windows console prompts are > #2031
+  [techtonik]
+
+- Expose Pillow package version as PIL.__version__ #2027
+  [techtonik]
+
 - Add Box and Hamming filters for resampling #1959
   [homm]
 
 - Retain a reference to core image object in PyAccess #2009
   [homm]
 
+3.3.1 (2016-08-18)
+------------------
+
+- Fix C90 compilation error for Tcl / Tk rewrite #2033
+  [matthew-brett]
+
+- Fix image loading when rotating by 0 deg #2052
+  [homm]
+
 3.3.0 (2016-07-01)
 ------------------
 

PIL/FpxImagePlugin.py

@@ -216,7 +216,7 @@ class FpxImageFile(ImageFile.ImageFile):
             self.fp = self.ole.openstream(self.stream[:2] +
                                           ["Subimage 0000 Data"])
 
-        ImageFile.ImageFile.load(self)
+        return ImageFile.ImageFile.load(self)
 
 #
 # --------------------------------------------------------------------

PIL/Image.py

@@ -246,7 +246,7 @@ else:
 
 _MODE_CONV = {
     # official modes
-    "1": ('|b1', None),  # broken
+    "1": ('|b1', None),  # Bits need to be extended to bytes
     "L": ('|u1', None),
     "LA": ('|u1', 2),
     "I": (_ENDIAN + 'i4', None),
@@ -615,17 +615,21 @@ class Image(object):
         self.save(b, 'PNG')
         return b.getvalue()
 
-    def __getattr__(self, name):
-        if name == "__array_interface__":
-            # numpy array interface support
-            new = {}
-            shape, typestr = _conv_type_shape(self)
-            new['shape'] = shape
-            new['typestr'] = typestr
+    @property
+    def __array_interface__(self):
+        # numpy array interface support
+        new = {}
+        shape, typestr = _conv_type_shape(self)
+        new['shape'] = shape
+        new['typestr'] = typestr
+        new['version'] = 3
+        if self.mode == '1':
+            # Binary images need to be extended from bits to bytes
+            # See: https://github.com/python-pillow/Pillow/issues/350
+            new['data'] = self.tobytes('raw', 'L')
+        else:
             new['data'] = self.tobytes()
-            new['version'] = 3
-            return new
-        raise AttributeError(name)
+        return new
 
     def __getstate__(self):
         return [
@@ -1877,7 +1881,7 @@ class Image(object):
             xs = float(x1 - x0) / w
             ys = float(y1 - y0) / h
             method = AFFINE
-            data = (xs, 0, x0 + xs/2, 0, ys, y0 + ys/2)
+            data = (xs, 0, x0, 0, ys, y0)
 
         elif method == PERSPECTIVE:
             data = data[0:8]

PIL/ImageOps.py

@@ -178,6 +178,27 @@ def crop(image, border=0):
         )
 
 
+def scale(image, factor, resample=Image.NEAREST):
+    """
+    Returns a rescaled image by a specific factor given in parameter.
+    A factor greater than 1 expands the image, between 0 and 1 contracts the
+    image.
+
+    :param factor: The expansion factor, as a float.
+    :param resample: An optional resampling filter. Same values possible as
+       in the PIL.Image.resize function.
+    :returns: An :py:class:`~PIL.Image.Image` object.
+    """
+    if factor == 1:
+        return image.copy()
+    elif factor <= 0:
+        raise ValueError("the factor must be greater than 0")
+    else:
+        size = (int(round(factor * image.width)),
+                int(round(factor * image.height)))
+        return image.resize(size, resample)
+
+
 def deform(image, deformer, resample=Image.BILINEAR):
     """
     Deform the image.

PIL/Jpeg2KImagePlugin.py

@@ -207,7 +207,7 @@ class Jpeg2KImageFile(ImageFile.ImageFile):
             t3 = (t[3][0], self.reduce, self.layers, t[3][3], t[3][4])
             self.tile = [(t[0], (0, 0) + self.size, t[2], t3)]
 
-        ImageFile.ImageFile.load(self)
+        return ImageFile.ImageFile.load(self)
 
 
 def _accept(prefix):

PIL/JpegImagePlugin.py

@@ -583,6 +583,14 @@ def _save(im, fp, filename):
     except KeyError:
         raise IOError("cannot write mode %s as JPEG" % im.mode)
 
+    if im.mode == 'RGBA':
+        warnings.warn(
+            'You are saving RGBA image as JPEG. The alpha channel will be '
+            'discarded. This conversion is deprecated and will be disabled '
+            'in Pillow 3.7. Please, convert the image to RGB explicitly.',
+            DeprecationWarning
+        )
+
     info = im.encoderinfo
 
     dpi = info.get("dpi", (0, 0))

PIL/__init__.py

@@ -14,6 +14,8 @@
 VERSION = '1.1.7'  # PIL version
 PILLOW_VERSION = '3.4.0.dev0'  # Pillow
 
+__version__ = PILLOW_VERSION
+
 _plugins = ['BmpImagePlugin',
             'BufrStubImagePlugin',
             'CurImagePlugin',

Tests/check_j2k_overflow.py

@@ -1,6 +1,7 @@
 from PIL import Image
 from helper import unittest, PillowTestCase
 
+
 class TestJ2kEncodeOverflow(PillowTestCase):
     def test_j2k_overflow(self):
 
@@ -12,7 +13,7 @@ class TestJ2kEncodeOverflow(PillowTestCase):
         except IOError as err:
             self.assertTrue(True, "IOError is expected")
         except Exception as err:
-            self.assertTrue(False, "Expected IOError, got %s" %type(err))
+            self.assertTrue(False, "Expected IOError, got %s" % type(err))
 
 if __name__ == '__main__':
     unittest.main()

Tests/helper.py

@@ -99,7 +99,7 @@ class PillowTestCase(unittest.TestCase):
             " average pixel value difference %.4f > epsilon %.4f" % (
                 ave_diff, epsilon))
 
-    def assert_warning(self, warn_class, func):
+    def assert_warning(self, warn_class, func, *args, **kwargs):
         import warnings
 
         result = None
@@ -108,7 +108,7 @@ class PillowTestCase(unittest.TestCase):
             warnings.simplefilter("always")
 
             # Hopefully trigger a warning.
-            result = func()
+            result = func(*args, **kwargs)
 
             # Verify some things.
             self.assertGreaterEqual(len(w), 1)

Tests/test_file_bmp.py

@@ -77,6 +77,5 @@ class TestFileBmp(PillowTestCase):
         self.assert_image_equal(im, target)
 
 
-
 if __name__ == '__main__':
     unittest.main()

Tests/test_file_jpeg.py

@@ -450,6 +450,25 @@ class TestFileJpeg(PillowTestCase):
         # Assert
         self.assertEqual(im.format, "JPEG")
 
+    def test_save_correct_modes(self):
+        out = BytesIO()
+        for mode in ['1', 'L', 'RGB', 'RGBX', 'CMYK', 'YCbCr']:
+            img = Image.new(mode, (20, 20))
+            img.save(out, "JPEG")
+
+    def test_save_wrong_modes(self):
+        out = BytesIO()
+        for mode in ['LA', 'La', 'RGBa', 'P']:
+            img = Image.new(mode, (20, 20))
+            self.assertRaises(IOError, img.save, out, "JPEG")
+
+    def test_save_modes_with_warnings(self):
+        # ref https://github.com/python-pillow/Pillow/issues/2005
+        out = BytesIO()
+        for mode in ['RGBA']:
+            img = Image.new(mode, (20, 20))
+            self.assert_warning(DeprecationWarning, img.save, out, "JPEG")
+
 
 if __name__ == '__main__':
     unittest.main()

Tests/test_file_jpeg2k.py

@@ -34,7 +34,8 @@ class TestFileJpeg2k(PillowTestCase):
         self.assertRegexpMatches(Image.core.jp2klib_version, '\d+\.\d+\.\d+$')
 
         im = Image.open('Tests/images/test-card-lossless.jp2')
-        im.load()
+        px = im.load()
+        self.assertEqual(px[0,0], (0, 0, 0))
         self.assertEqual(im.mode, 'RGB')
         self.assertEqual(im.size, (640, 480))
         self.assertEqual(im.format, 'JPEG2000')

Tests/test_file_libtiff.py

@@ -170,7 +170,8 @@ class TestFileLibTiff(LibTiffTestCase):
                                 'RowsPerStrip',
                                 'StripOffsets']
             for field in requested_fields:
-                self.assertTrue(field in reloaded, "%s not in metadata" % field)
+                self.assertTrue(field in reloaded,
+                                "%s not in metadata" % field)
 
     def test_additional_metadata(self):
         # these should not crash. Seriously dummy data, most of it doesn't make
@@ -183,7 +184,8 @@ class TestFileLibTiff(LibTiffTestCase):
                                                         in TiffTags.LIBTIFF_CORE]
                           if info.type is not None)
 
-        # Exclude ones that have special meaning that we're already testing them
+        # Exclude ones that have special meaning
+        # that we're already testing them
         im = Image.open('Tests/images/hopper_g4.tif')
         for tag in im.tag_v2.keys():
             try:
@@ -422,8 +424,8 @@ class TestFileLibTiff(LibTiffTestCase):
     def test_gray_semibyte_per_pixel(self):
         test_files = (
             (
-                24.8,#epsilon
-                (#group
+                24.8,  # epsilon
+                (  # group
                     "Tests/images/tiff_gray_2_4_bpp/hopper2.tif",
                     "Tests/images/tiff_gray_2_4_bpp/hopper2I.tif",
                     "Tests/images/tiff_gray_2_4_bpp/hopper2R.tif",
@@ -431,8 +433,8 @@ class TestFileLibTiff(LibTiffTestCase):
                 )
             ),
             (
-                7.3,#epsilon
-                (#group
+                7.3,  # epsilon
+                (  # group
                     "Tests/images/tiff_gray_2_4_bpp/hopper4.tif",
                     "Tests/images/tiff_gray_2_4_bpp/hopper4I.tif",
                     "Tests/images/tiff_gray_2_4_bpp/hopper4R.tif",
@@ -504,6 +506,5 @@ class TestFileLibTiff(LibTiffTestCase):
         im.save(outfile)
 
 
-
 if __name__ == '__main__':
     unittest.main()

Tests/test_file_pcx.py

@@ -1,6 +1,6 @@
 from helper import unittest, PillowTestCase, hopper
 
-from PIL import Image, PcxImagePlugin
+from PIL import Image, ImageFile, PcxImagePlugin
 
 
 class TestFilePcx(PillowTestCase):
@@ -46,6 +46,84 @@ class TestFilePcx(PillowTestCase):
         # Make sure all pixels are either 0 or 255.
         self.assertEqual(im.histogram()[0] + im.histogram()[255], 447*144)
 
+    def test_1px_width(self):
+        im = Image.new('L', (1, 256))
+        px = im.load()
+        for y in range(256):
+            px[0, y] = y
+        self._roundtrip(im)
+
+    def test_large_count(self):
+        im = Image.new('L', (256, 1))
+        px = im.load()
+        for x in range(256):
+            px[x, 0] = x // 67 * 67
+        self._roundtrip(im)
+
+    def _test_buffer_overflow(self, im, size=1024):
+        _last = ImageFile.MAXBLOCK
+        ImageFile.MAXBLOCK = size
+        try:
+            self._roundtrip(im)
+        finally:
+            ImageFile.MAXBLOCK = _last
+
+    def test_break_in_count_overflow(self):
+        im = Image.new('L', (256, 5))
+        px = im.load()
+        for y in range(4):
+            for x in range(256):
+                px[x, y] = x % 128
+        self._test_buffer_overflow(im)
+
+    def test_break_one_in_loop(self):
+        im = Image.new('L', (256, 5))
+        px = im.load()
+        for y in range(5):
+            for x in range(256):
+                px[x, y] = x % 128
+        self._test_buffer_overflow(im)
+
+    def test_break_many_in_loop(self):
+        im = Image.new('L', (256, 5))
+        px = im.load()
+        for y in range(4):
+            for x in range(256):
+                px[x, y] = x % 128
+        for x in range(8):
+            px[x, 4] = 16
+        self._test_buffer_overflow(im)
+
+    def test_break_one_at_end(self):
+        im = Image.new('L', (256, 5))
+        px = im.load()
+        for y in range(5):
+            for x in range(256):
+                px[x, y] = x % 128
+        px[0, 3] = 128 + 64
+        self._test_buffer_overflow(im)
+
+    def test_break_many_at_end(self):
+        im = Image.new('L', (256, 5))
+        px = im.load()
+        for y in range(5):
+            for x in range(256):
+                px[x, y] = x % 128
+        for x in range(4):
+            px[x * 2, 3] = 128 + 64
+            px[x + 256 - 4, 3] = 0
+        self._test_buffer_overflow(im)
+
+    def test_break_padding(self):
+        im = Image.new('L', (257, 5))
+        px = im.load()
+        for y in range(5):
+            for x in range(257):
+                px[x, y] = x % 128
+        for x in range(5):
+            px[x, 3] = 0
+        self._test_buffer_overflow(im)
+
 
 if __name__ == '__main__':
     unittest.main()

Tests/test_file_png.py

@@ -313,7 +313,7 @@ class TestFilePng(PillowTestCase):
         #                  -14: malformed chunk
 
         for offset in (-10, -13, -14):
-            with open(TEST_PNG_FILE,'rb') as f:
+            with open(TEST_PNG_FILE, 'rb') as f:
                 test_file = f.read()[:offset]
 
             im = Image.open(BytesIO(test_file))
@@ -347,7 +347,6 @@ class TestFilePng(PillowTestCase):
         finally:
             ImageFile.LOAD_TRUNCATED_IMAGES = False
 
-
     def test_roundtrip_dpi(self):
         # Check dpi roundtripping
 

Tests/test_image_access.py

@@ -105,6 +105,7 @@ class TestCffiPutPixel(TestImagePutPixel):
     def setUp(self):
         try:
             import cffi
+            assert cffi  # silence warning
         except ImportError:
             self.skipTest("No cffi")
 
@@ -115,6 +116,7 @@ class TestCffiGetPixel(TestImageGetPixel):
     def setUp(self):
         try:
             import cffi
+            assert cffi  # silence warning
         except ImportError:
             self.skipTest("No cffi")
 
@@ -125,6 +127,7 @@ class TestCffi(AccessTest):
     def setUp(self):
         try:
             import cffi
+            assert cffi  # silence warning
         except ImportError:
             self.skipTest("No cffi")
 
@@ -229,7 +232,7 @@ class TestCffi(AccessTest):
             # Do not save references to the image, only to the access object
             px = Image.new('L', (size, 1), 0).load()
             for i in range(size):
-                # pixels can contain garbarge if image is released
+                # pixels can contain garbage if image is released
                 self.assertEqual(px[i, 0], 0)
 
 

Tests/test_image_array.py

@@ -25,13 +25,26 @@ class TestImageArray(PillowTestCase):
         self.assertEqual(test("RGBX"), (3, (100, 128, 4), '|u1', 51200))
 
     def test_fromarray(self):
+
+        class Wrapper(object):
+            """ Class with API matching Image.fromarray """
+
+            def __init__(self, img, arr_params):
+                self.img = img
+                self.__array_interface__ = arr_params
+
+            def tobytes(self):
+                return self.img.tobytes()
+
         def test(mode):
             i = im.convert(mode)
             a = i.__array_interface__
             a["strides"] = 1  # pretend it's non-contigous
-            i.__array_interface__ = a  # patch in new version of attribute
-            out = Image.fromarray(i)
+            # Make wrapper instance for image, new array interface
+            wrapped = Wrapper(i, a)
+            out = Image.fromarray(wrapped)
             return out.mode, out.size, list(i.getdata()) == list(out.getdata())
+
         # self.assertEqual(test("1"), ("1", (128, 100), True))
         self.assertEqual(test("L"), ("L", (128, 100), True))
         self.assertEqual(test("I"), ("I", (128, 100), True))

Tests/test_image_getdata.py

@@ -20,13 +20,13 @@ class TestImageGetData(PillowTestCase):
             return data[0], len(data), len(list(data))
 
         self.assertEqual(getdata("1"), (0, 960, 960))
-        self.assertEqual(getdata("L"), (25, 960, 960))
-        self.assertEqual(getdata("I"), (25, 960, 960))
-        self.assertEqual(getdata("F"), (25.0, 960, 960))
-        self.assertEqual(getdata("RGB"), (((20, 20, 70), 960, 960)))
-        self.assertEqual(getdata("RGBA"), ((20, 20, 70, 255), 960, 960))
-        self.assertEqual(getdata("CMYK"), ((235, 235, 185, 0), 960, 960))
-        self.assertEqual(getdata("YCbCr"), ((25, 153, 123), 960, 960))
+        self.assertEqual(getdata("L"), (16, 960, 960))
+        self.assertEqual(getdata("I"), (16, 960, 960))
+        self.assertEqual(getdata("F"), (16.0, 960, 960))
+        self.assertEqual(getdata("RGB"), (((11, 13, 52), 960, 960)))
+        self.assertEqual(getdata("RGBA"), ((11, 13, 52, 255), 960, 960))
+        self.assertEqual(getdata("CMYK"), ((244, 242, 203, 0), 960, 960))
+        self.assertEqual(getdata("YCbCr"), ((16, 147, 123), 960, 960))
 
 
 if __name__ == '__main__':

Tests/test_image_resample.py

@@ -320,5 +320,22 @@ class CoreResamplePassesTest(PillowTestCase):
         self.assertEqual(Image.core.getcount(), count + 2)
 
 
+class CoreResampleCoefficientsTest(PillowTestCase):
+    def test_reduce(self):
+        test_color = 254
+        # print ''
+
+        for size in range(400000, 400010, 2):
+            # print '\r', size,
+            i = Image.new('L', (size, 1), 0)
+            draw = ImageDraw.Draw(i)
+            draw.rectangle((0, 0, i.size[0] // 2 - 1, 0), test_color)
+
+            px = i.resize((5, i.size[1]), Image.BICUBIC).load()
+            if px[2, 0] != test_color // 2:
+                self.assertEqual(test_color // 2, px[2, 0])
+                # print '\r>', size, test_color // 2, px[2, 0]
+
+
 if __name__ == '__main__':
     unittest.main()

Tests/test_image_transform.py

@@ -1,3 +1,5 @@
+import math
+
 from helper import unittest, PillowTestCase, hopper
 
 from PIL import Image
@@ -98,7 +100,7 @@ class TestImageTransform(PillowTestCase):
     def test_alpha_premult_resize(self):
 
         def op(im, sz):
-            return im.resize(sz, Image.LINEAR)
+            return im.resize(sz, Image.BILINEAR)
 
         self._test_alpha_premult(op)
 
@@ -139,5 +141,117 @@ class TestImageTransform(PillowTestCase):
         self.test_mesh()
 
 
+class TestImageTransformAffine(PillowTestCase):
+    transform = Image.AFFINE
+
+    def _test_image(self):
+        im = hopper('RGB')
+        return im.crop((10, 20, im.width - 10, im.height - 20))
+
+    def _test_rotate(self, deg, transpose):
+        im = self._test_image()
+
+        angle = - math.radians(deg)
+        matrix = [
+            round(math.cos(angle), 15), round(math.sin(angle), 15), 0.0,
+            round(-math.sin(angle), 15), round(math.cos(angle), 15), 0.0,
+            0, 0]
+        matrix[2] = (1 - matrix[0] - matrix[1]) * im.width / 2
+        matrix[5] = (1 - matrix[3] - matrix[4]) * im.height / 2
+
+        if transpose is not None:
+            transposed = im.transpose(transpose)
+        else:
+            transposed = im
+
+        for resample in [Image.NEAREST, Image.BILINEAR, Image.BICUBIC]:
+            transformed = im.transform(transposed.size, self.transform,
+                                       matrix, resample)
+            self.assert_image_equal(transposed, transformed)
+
+    def test_rotate_0_deg(self):
+        self._test_rotate(0, None)
+
+    def test_rotate_90_deg(self):
+        self._test_rotate(90, Image.ROTATE_90)
+
+    def test_rotate_180_deg(self):
+        self._test_rotate(180, Image.ROTATE_180)
+
+    def test_rotate_270_deg(self):
+        self._test_rotate(270, Image.ROTATE_270)
+
+    def _test_resize(self, scale, epsilonscale):
+        im = self._test_image()
+
+        size_up = int(round(im.width * scale)), int(round(im.height * scale))
+        matrix_up = [
+            1 / scale, 0, 0,
+            0, 1 / scale, 0,
+            0, 0]
+        matrix_down = [
+            scale, 0, 0,
+            0, scale, 0,
+            0, 0]
+
+        for resample, epsilon in [(Image.NEAREST, 0),
+                                  (Image.BILINEAR, 2), (Image.BICUBIC, 1)]:
+            transformed = im.transform(
+                size_up, self.transform, matrix_up, resample)
+            transformed = transformed.transform(
+                im.size, self.transform, matrix_down, resample)
+            self.assert_image_similar(transformed, im, epsilon * epsilonscale)
+
+    def test_resize_1_1x(self):
+        self._test_resize(1.1, 6.9)
+
+    def test_resize_1_5x(self):
+        self._test_resize(1.5, 5.5)
+
+    def test_resize_2_0x(self):
+        self._test_resize(2.0, 5.5)
+
+    def test_resize_2_3x(self):
+        self._test_resize(2.3, 3.7)
+
+    def test_resize_2_5x(self):
+        self._test_resize(2.5, 3.7)
+
+    def _test_translate(self, x, y, epsilonscale):
+        im = self._test_image()
+
+        size_up = int(round(im.width + x)), int(round(im.height + y))
+        matrix_up = [
+            1, 0, -x,
+            0, 1, -y,
+            0, 0]
+        matrix_down = [
+            1, 0, x,
+            0, 1, y,
+            0, 0]
+
+        for resample, epsilon in [(Image.NEAREST, 0),
+                                  (Image.BILINEAR, 1.5), (Image.BICUBIC, 1)]:
+            transformed = im.transform(
+                size_up, self.transform, matrix_up, resample)
+            transformed = transformed.transform(
+                im.size, self.transform, matrix_down, resample)
+            self.assert_image_similar(transformed, im, epsilon * epsilonscale)
+
+    def test_translate_0_1(self):
+        self._test_translate(.1, 0, 3.7)
+
+    def test_translate_0_6(self):
+        self._test_translate(.6, 0, 9.1)
+
+    def test_translate_50(self):
+        self._test_translate(50, 50, 0)
+
+
+class TestImageTransformPerspective(TestImageTransformAffine):
+    # Repeat all tests for AFFINE transormations with PERSPECTIVE
+    transform = Image.PERSPECTIVE
+
+
 if __name__ == '__main__':
     unittest.main()

Tests/test_imageops.py

@@ -78,6 +78,22 @@ class TestImageOps(PillowTestCase):
         ImageOps.equalize(i.convert("P"))
         ImageOps.equalize(i.convert("RGB"))
 
+    def test_scale(self):
+        # Test the scaling function
+        i = hopper("L").resize((50, 50))
+        
+        with self.assertRaises(ValueError):
+            ImageOps.scale(i, -1)
+
+        newimg = ImageOps.scale(i, 1)
+        self.assertEqual(newimg.size, (50, 50))
+
+        newimg = ImageOps.scale(i, 2)
+        self.assertEqual(newimg.size, (100, 100))
+
+        newimg = ImageOps.scale(i, 0.5)
+        self.assertEqual(newimg.size, (25, 25))
+
 
 if __name__ == '__main__':
     unittest.main()

Tests/test_imagepath.py

@@ -5,6 +5,7 @@ from PIL import ImagePath, Image
 import array
 import struct
 
+
 class TestImagePath(PillowTestCase):
 
     def test_path(self):
@@ -61,7 +62,6 @@ class TestImagePath(PillowTestCase):
             p = ImagePath.Path(arr.tostring())
         self.assertEqual(list(p), [(0.0, 1.0)])
 
-
     def test_overflow_segfault(self):
         try:
             # post patch, this fails with a memory error

Tests/test_imagesequence.py

@@ -56,7 +56,7 @@ class TestImageSequence(PillowTestCase):
         im = Image.open('Tests/images/multipage.tiff')
         firstFrame = None
         for frame in ImageSequence.Iterator(im):
-            if firstFrame == None:
+            if firstFrame is None:
                 firstFrame = frame.copy()
             pass
         for frame in ImageSequence.Iterator(im):

Tests/test_imagetk.py

@@ -8,6 +8,7 @@ except (OSError, ImportError) as v:
     # Skipped via setUp()
     pass
 
+
 class TestImageTk(PillowTestCase):
 
     def setUp(self):
@@ -24,7 +25,7 @@ class TestImageTk(PillowTestCase):
         im2 = Image.open(TEST_PNG)
         with open(TEST_PNG, 'rb') as fp:
             data = fp.read()
-        kw = {"file":TEST_JPG, "data":data}
+        kw = {"file": TEST_JPG, "data": data}
 
         # Test "file"
         im = ImageTk._get_image_from_kw(kw)

Tests/test_imagewin_pointers.py

@@ -57,7 +57,10 @@ if sys.platform.startswith('win32'):
     DeleteObject.argtypes = [ctypes.wintypes.HGDIOBJ]
 
     CreateDIBSection = ctypes.windll.gdi32.CreateDIBSection
-    CreateDIBSection.argtypes = [ctypes.wintypes.HDC, ctypes.c_void_p, ctypes.c_uint, ctypes.POINTER(ctypes.c_void_p), ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD]
+    CreateDIBSection.argtypes = [ctypes.wintypes.HDC, ctypes.c_void_p,
+                                 ctypes.c_uint,
+                                 ctypes.POINTER(ctypes.c_void_p),
+                                 ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD]
     CreateDIBSection.restype = ctypes.wintypes.HBITMAP
 
     def serialize_dib(bi, pixels):
@@ -99,7 +102,8 @@ if sys.platform.startswith('win32'):
             hdc = CreateCompatibleDC(None)
             # print('hdc:',hex(hdc))
             pixels = ctypes.c_void_p()
-            dib = CreateDIBSection(hdc, ctypes.byref(hdr), DIB_RGB_COLORS, ctypes.byref(pixels), None, 0)
+            dib = CreateDIBSection(hdc, ctypes.byref(hdr), DIB_RGB_COLORS,
+                                   ctypes.byref(pixels), None, 0)
             SelectObject(hdc, dib)
 
             imdib.expose(hdc)

Tests/test_lib_pack.py

@@ -10,11 +10,11 @@ class TestLibPack(PillowTestCase):
 
     def test_pack(self):
 
-        def pack(mode, rawmode):
+        def pack(mode, rawmode, in_data=(1, 2, 3, 4)):
             if len(mode) == 1:
-                im = Image.new(mode, (1, 1), 1)
+                im = Image.new(mode, (1, 1), in_data[0])
             else:
-                im = Image.new(mode, (1, 1), (1, 2, 3, 4)[:len(mode)])
+                im = Image.new(mode, (1, 1), in_data[:len(mode)])
 
             if py3:
                 return list(im.tobytes("raw", rawmode))
@@ -47,6 +47,8 @@ class TestLibPack(PillowTestCase):
         self.assertEqual(pack("RGBX", "RGBX"), [1, 2, 3, 4])  # 4->255?
 
         self.assertEqual(pack("RGBA", "RGBA"), [1, 2, 3, 4])
+        self.assertEqual(pack("RGBA", "BGRa"), [0, 0, 0, 4])
+        self.assertEqual(pack("RGBA", "BGRa", in_data=(20, 30, 40, 50)), [8, 6, 4, 50])
 
         self.assertEqual(pack("RGBa", "RGBa"), [1, 2, 3, 4])
         self.assertEqual(pack("RGBa", "BGRa"), [3, 2, 1, 4])
@@ -122,7 +124,9 @@ class TestLibPack(PillowTestCase):
         self.assertEqual(unpack("RGB", "XBGR", 4), (4, 3, 2))
 
         self.assertEqual(unpack("RGBA", "RGBA", 4), (1, 2, 3, 4))
+        self.assertEqual(unpack("RGBA", "RGBa", 4), (63, 127, 191, 4))
         self.assertEqual(unpack("RGBA", "BGRA", 4), (3, 2, 1, 4))
+        self.assertEqual(unpack("RGBA", "BGRa", 4), (191, 127, 63, 4))
         self.assertEqual(unpack("RGBA", "ARGB", 4), (2, 3, 4, 1))
         self.assertEqual(unpack("RGBA", "ABGR", 4), (4, 3, 2, 1))
         self.assertEqual(unpack("RGBA", "RGBA;15", 2), (8, 131, 0, 0))

Tests/test_numpy.py

@@ -1,4 +1,5 @@
 from __future__ import print_function
+import sys
 from helper import unittest, PillowTestCase, hopper
 
 from PIL import Image
@@ -6,18 +7,29 @@ from PIL import Image
 try:
     import site
     import numpy
+    assert site  # silence warning
+    assert numpy  # silence warning
 except ImportError:
     # Skip via setUp()
     pass
 
 TEST_IMAGE_SIZE = (10, 10)
 
+# Numpy on pypy as of pypy 5.3.1 is corrupting the numpy.array(Image)
+# call such that it's returning a object of type numpy.ndarray, but
+# the repr is that of a PIL.Image. Size and shape are 1 and (), not the
+# size and shape of the array. This causes failures in several tests.
+SKIP_NUMPY_ON_PYPY = hasattr(sys, 'pypy_version_info') and (
+    sys.pypy_version_info <= (5,3,1,'final',0))
+
 class TestNumpy(PillowTestCase):
 
     def setUp(self):
         try:
             import site
             import numpy
+            assert site  # silence warning
+            assert numpy  # silence warning
         except ImportError:
             self.skipTest("ImportError")
 
@@ -100,6 +112,7 @@ class TestNumpy(PillowTestCase):
         self.assert_image(Image.fromarray(a[:, :, 1]), "L", TEST_IMAGE_SIZE)
 
     def _test_img_equals_nparray(self, img, np):
+        self.assertGreaterEqual(len(np.shape), 2)
         np_size = np.shape[1], np.shape[0]
         self.assertEqual(img.size, np_size)
         px = img.load()
@@ -107,12 +120,22 @@ class TestNumpy(PillowTestCase):
             for y in range(0, img.size[1], int(img.size[1]/10)):
                 self.assert_deep_equal(px[x, y], np[y, x])
 
+    @unittest.skipIf(SKIP_NUMPY_ON_PYPY, "numpy.array(Image) is flaky on PyPy")
     def test_16bit(self):
         img = Image.open('Tests/images/16bit.cropped.tif')
         np_img = numpy.array(img)
         self._test_img_equals_nparray(img, np_img)
         self.assertEqual(np_img.dtype, numpy.dtype('<u2'))
 
+    def test_1bit(self):
+        # Test that 1-bit arrays convert to numpy and back
+        # See: https://github.com/python-pillow/Pillow/issues/350
+        arr = numpy.array([[1, 0, 0, 1, 0], [0, 1, 0, 0, 0]], 'u1')
+        img = Image.fromarray(arr * 255).convert('1')
+        self.assertEqual(img.mode, '1')
+        arr_back = numpy.array(img)
+        numpy.testing.assert_array_equal(arr, arr_back)
+
     def test_save_tiff_uint16(self):
         '''
         Open a single-channel uint16 greyscale image and verify that it can be saved without
@@ -128,6 +151,7 @@ class TestNumpy(PillowTestCase):
         im_good = Image.open(filename)
         self.assert_image_equal(im_good, im_test)
 
+    @unittest.skipIf(SKIP_NUMPY_ON_PYPY, "numpy.array(Image) is flaky on PyPy")
     def test_to_array(self):
 
         def _to_array(mode, dtype):

Tk/tkImaging.c

@@ -371,7 +371,8 @@ int load_tkinter_funcs(void)
 #if PY_VERSION_HEX >= 0x03000000
 char *fname2char(PyObject *fname)
 {
-    PyObject *bytes = PyUnicode_EncodeFSDefault(fname);
+    PyObject* bytes;
+    bytes = PyUnicode_EncodeFSDefault(fname);
     if (bytes == NULL) {
         return NULL;
     }
@@ -391,9 +392,10 @@ void *_dfunc(void *lib_handle, const char *func_name)
      * Returns function pointer or NULL if not present.
      */
 
+    void* func;
     /* Reset errors. */
     dlerror();
-    void *func = dlsym(lib_handle, func_name);
+    func = dlsym(lib_handle, func_name);
     if (func == NULL) {
         const char *error = dlerror();
         PyErr_SetString(PyExc_RuntimeError, error);

_imagingcms.c

@@ -729,7 +729,7 @@ static cmsBool _calculate_rgb_primaries(CmsProfileObject* self, cmsCIEXYZTRIPLE*
 
     /* http://littlecms2.blogspot.com/2009/07/less-is-more.html */
 
-    // double array of RGB values with max on each identitiy
+    // double array of RGB values with max on each identity
     hXYZ = cmsCreateXYZProfile();
     if (hXYZ == NULL)
         return 0;

decode.c

@@ -204,7 +204,7 @@ _setimage(ImagingDecoderObject* decoder, PyObject* args)
             }
             state->bytes = (state->bits * state->xsize+7)/8;
         }
-        /* malloc check ok, oveflow checked above */
+        /* malloc check ok, overflow checked above */
         state->buffer = (UINT8*) malloc(state->bytes);
         if (!state->buffer)
             return PyErr_NoMemory();

docs/handbook/image-file-formats.rst

@@ -16,6 +16,8 @@ explicitly.
 Fully supported formats
 -----------------------
 
+.. contents:: 
+
 BMP
 ^^^
 

docs/installation.rst

@@ -51,11 +51,11 @@ supported Pythons in both 32 and 64-bit versions in wheel, egg, and
 executable installers. These binaries have all of the optional
 libraries included::
 
-  $ pip install Pillow
+  > pip install Pillow
 
 or::
 
-  $ easy_install Pillow
+  > easy_install Pillow
 
 
 OS X Installation
@@ -282,7 +282,9 @@ Building on Linux
 ^^^^^^^^^^^^^^^^^
 
 If you didn't build Python from source, make sure you have Python's
-development libraries installed. In Debian or Ubuntu::
+development libraries installed. 
+
+In Debian or Ubuntu::
 
     $ sudo apt-get install python-dev python-setuptools
 
@@ -294,6 +296,10 @@ In Fedora, the command is::
 
     $ sudo dnf install python-devel redhat-rpm-config
     
+Or for Python 3::
+
+    $ sudo dnf install python3-devel redhat-rpm-config 
+
 .. Note:: ``redhat-rpm-config`` is required on Fedora 23, but not earlier versions.
 
 Prerequisites are installed on **Ubuntu 12.04 LTS** or **Raspian Wheezy

docs/releasenotes/3.4.0.rst

@@ -0,0 +1,15 @@
+
+3.4.0
+-----
+
+New resizing filters
+====================
+
+Two new filters available for ``Image.resize()`` and ``Image.thumbnail()``
+functions: ``BOX`` and ``HAMMING``. ``BOX`` is the high-performance filter with
+two times shorter window than ``BILINEAR``. It can be used for image reduction
+3 and more times and produces a more sharp result than ``BILINEAR``.
+
+``HAMMING`` filter has the same performance as ``BILINEAR`` filter while
+providing the image downscaling quality comparable to ``BICUBIC``.
+Both new filters don't show good quality for the image upscaling.

docs/releasenotes/index.rst

@@ -6,6 +6,7 @@ Release Notes
 .. toctree::
   :maxdepth: 2
 
+  3.4.0
   3.3.0
   3.2.0
   3.1.2

encode.c

@@ -365,7 +365,7 @@ get_packer(ImagingEncoderObject* encoder, const char* mode,
     pack = ImagingFindPacker(mode, rawmode, &bits);
     if (!pack) {
         Py_DECREF(encoder);
-        PyErr_SetString(PyExc_SystemError, "unknown raw mode");
+        PyErr_Format(PyExc_ValueError, "No packer found from %s to %s", mode, rawmode);
         return -1;
     }
 

libImaging/Geometry.c

@@ -240,7 +240,7 @@ ImagingRotate270(Imaging imOut, Imaging imIn)
 /* transform primitives (ImagingTransformMap) */
 
 static int
-affine_transform(double* xin, double* yin, int x, int y, void* data)
+affine_transform(double* xout, double* yout, int x, int y, void* data)
 {
     /* full moon tonight.  your compiler will generate bogus code
        for simple expressions, unless you reorganize the code, or
@@ -250,28 +250,34 @@ affine_transform(double* xin, double* yin, int x, int y, void* data)
     double a0 = a[0]; double a1 = a[1]; double a2 = a[2];
     double a3 = a[3]; double a4 = a[4]; double a5 = a[5];
 
-    xin[0] = a0*x + a1*y + a2;
-    yin[0] = a3*x + a4*y + a5;
+    double xin = x + 0.5;
+    double yin = y + 0.5;
+
+    xout[0] = a0*xin + a1*yin + a2;
+    yout[0] = a3*xin + a4*yin + a5;
 
     return 1;
 }
 
 static int
-perspective_transform(double* xin, double* yin, int x, int y, void* data)
+perspective_transform(double* xout, double* yout, int x, int y, void* data)
 {
     double* a = (double*) data;
     double a0 = a[0]; double a1 = a[1]; double a2 = a[2];
     double a3 = a[3]; double a4 = a[4]; double a5 = a[5];
     double a6 = a[6]; double a7 = a[7];
 
-    xin[0] = (a0*x + a1*y + a2) / (a6*x + a7*y + 1);
-    yin[0] = (a3*x + a4*y + a5) / (a6*x + a7*y + 1);
+    double xin = x + 0.5;
+    double yin = y + 0.5;
+
+    xout[0] = (a0*xin + a1*yin + a2) / (a6*xin + a7*yin + 1);
+    yout[0] = (a3*xin + a4*yin + a5) / (a6*xin + a7*yin + 1);
 
     return 1;
 }
 
 static int
-quad_transform(double* xin, double* yin, int x, int y, void* data)
+quad_transform(double* xout, double* yout, int x, int y, void* data)
 {
     /* quad warp: map quadrilateral to rectangle */
 
@@ -279,8 +285,11 @@ quad_transform(double* xin, double* yin, int x, int y, void* data)
     double a0 = a[0]; double a1 = a[1]; double a2 = a[2]; double a3 = a[3];
     double a4 = a[4]; double a5 = a[5]; double a6 = a[6]; double a7 = a[7];
 
-    xin[0] = a0 + a1*x + a2*y + a3*x*y;
-    yin[0] = a4 + a5*x + a6*y + a7*x*y;
+    double xin = x + 0.5;
+    double yin = y + 0.5;
+
+    xout[0] = a0 + a1*xin + a2*yin + a3*xin*yin;
+    yout[0] = a4 + a5*xin + a6*yin + a7*xin*yin;
 
     return 1;
 }
@@ -691,8 +700,8 @@ ImagingScaleAffine(Imaging imOut, Imaging imIn,
         return (Imaging) ImagingError_MemoryError();
     }
 
-    xo = a[2];
-    yo = a[5];
+    xo = a[2] + a[0] * 0.5;
+    yo = a[5] + a[4] * 0.5;
 
     xmin = x1;
     xmax = x0;
@@ -771,8 +780,10 @@ affine_fixed(Imaging imOut, Imaging imIn,
 /* use 16.16 fixed point arithmetics */
 #define FIX(v) FLOOR((v)*65536.0 + 0.5)
 
-    a0 = FIX(a[0]); a1 = FIX(a[1]); a2 = FIX(a[2]);
-    a3 = FIX(a[3]); a4 = FIX(a[4]); a5 = FIX(a[5]);
+    a0 = FIX(a[0]); a1 = FIX(a[1]); 
+    a3 = FIX(a[3]); a4 = FIX(a[4]);
+    a2 = FIX(a[2] + a[0] * 0.5 + a[1] * 0.5);
+    a5 = FIX(a[5] + a[3] * 0.5 + a[4] * 0.5);
 
 #undef FIX
 
@@ -835,9 +846,10 @@ ImagingTransformAffine(Imaging imOut, Imaging imIn,
             filterid, fill);
     }
 
-    if (a[1] == 0 && a[3] == 0)
+    if (a[1] == 0 && a[3] == 0) {
         /* Scaling */
         return ImagingScaleAffine(imOut, imIn, x0, y0, x1, y1, a, fill);
+    }
 
     if (!imOut || !imIn || strcmp(imIn->mode, imOut->mode) != 0)
         return (Imaging) ImagingError_ModeError();
@@ -867,8 +879,8 @@ ImagingTransformAffine(Imaging imOut, Imaging imIn,
     xsize = (int) imIn->xsize;
     ysize = (int) imIn->ysize;
 
-    xo = a[2];
-    yo = a[5];
+    xo = a[2] + a[1] * 0.5 + a[0] * 0.5;
+    yo = a[5] + a[4] * 0.5 + a[3] * 0.5;
 
 #define AFFINE_TRANSFORM(pixel, image)\
     for (y = y0; y < y1; y++) {\

libImaging/Pack.c

@@ -72,6 +72,10 @@
 #define C64L C64N
 #endif
 
+/* like (a * b + 127) / 255), but much faster on most platforms */
+#define MULDIV255(a, b, tmp)\
+        (tmp = (a) * (b) + 128, ((((tmp) >> 8) + (tmp)) >> 8))
+
 static void
 pack1(UINT8* out, const UINT8* in, int pixels)
 {
@@ -315,6 +319,21 @@ ImagingPackABGR(UINT8* out, const UINT8* in, int pixels)
     }
 }
 
+void
+ImagingPackBGRa(UINT8* out, const UINT8* in, int pixels)
+{
+    int i;
+    /* BGRa, reversed bytes with premultiplied alpha */
+    for (i = 0; i < pixels; i++) {
+        int alpha = out[3] = in[A];
+        int tmp;
+        out[0] = MULDIV255(in[B], alpha, tmp);
+        out[1] = MULDIV255(in[G], alpha, tmp);
+        out[2] = MULDIV255(in[R], alpha, tmp);
+        out += 4; in += 4;
+    }
+}
+
 static void
 packRGBL(UINT8* out, const UINT8* in, int pixels)
 {
@@ -525,6 +544,7 @@ static struct {
     {"RGBA",	"BGR",		24,	ImagingPackBGR},
     {"RGBA",	"BGRA",		32,	ImagingPackBGRA},
     {"RGBA",	"ABGR",		32,	ImagingPackABGR},
+    {"RGBA",	"BGRa",		32,	ImagingPackBGRa},
     {"RGBA",   	"R",            8,      band0},
     {"RGBA",   	"G",            8,      band1},
     {"RGBA",   	"B",            8,      band2},

libImaging/PcxEncode.c

@@ -75,7 +75,7 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
                            (UINT8*) im->image[state->y + state->yoff] +
                            state->xoff * im->pixelsize, state->xsize);
 
-            state->y++;
+            state->y += 1;
 
             state->count = 1;
             state->LAST = state->buffer[0];
@@ -91,7 +91,7 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
             /* when we arrive here, "count" contains the number of
                bytes having the value of "LAST" that we've already
                seen */
-            while (state->x < planes * bytes_per_line) {
+            do {
                 /* If we're encoding an odd width file, and we've
                    got more than one plane, we need to pad each
                    color row with padding bytes at the end. Since
@@ -103,22 +103,23 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
 
                     if (state->count == 63) {
                         /* this run is full; flush it */
-                        if (bytes < 2)
+                        if (bytes < 2) {
                             return ptr - buf;
-                        *ptr++ = 0xff;
-                        *ptr++ = state->LAST;
+                        }
+                        ptr[0] = 0xff;
+                        ptr[1] = state->LAST;
+                        ptr += 2;
                         bytes -= 2;
 
                         state->count = 0;
-
                     }
 
                     this = state->buffer[state->x];
 
                     if (this == state->LAST) {
                         /* extend the current run */
-                        state->x++;
-                        state->count++;
+                        state->x += 1;
+                        state->count += 1;
 
                     } else {
                         /* start a new run */
@@ -126,15 +127,17 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
                             if (bytes < 1) {
                                 return ptr - buf;
                             }
-                            *ptr++ = state->LAST;
-                            bytes--;
+                            ptr[0] = state->LAST;
+                            ptr += 1;
+                            bytes -= 1;
                         } else {
                             if (state->count > 0) {
                                 if (bytes < 2) {
                                     return ptr - buf;
                                 }
-                                *ptr++ = 0xc0 | state->count;
-                                *ptr++ = state->LAST;
+                                ptr[0] = 0xc0 | state->count;
+                                ptr[1] = state->LAST;
+                                ptr += 2;
                                 bytes -= 2;
                             }
                         }
@@ -142,8 +145,7 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
                         state->LAST = this;
                         state->count = 1;
 
-                        state->x++;
-
+                        state->x += 1;
                     }
                 }
 
@@ -152,33 +154,34 @@ ImagingPcxEncode(Imaging im, ImagingCodecState state, UINT8* buf, int bytes)
                     if (bytes < 1 + padding) {
                         return ptr - buf;
                     }
-                    *ptr++ = state->LAST;
-                    bytes--;
+                    ptr[0] = state->LAST;
+                    ptr += 1;
+                    bytes -= 1;
                 } else {
                     if (state->count > 0) {
                         if (bytes < 2 + padding) {
                             return ptr - buf;
                         }
-                        *ptr++ = 0xc0 | state->count;
-                        *ptr++ = state->LAST;
+                        ptr[0] = 0xc0 | state->count;
+                        ptr[1] = state->LAST;
+                        ptr += 2;
                         bytes -= 2;
                     }
                 }
-                if (bytes < padding) {
-                    return ptr - buf;
-                }
                 /* add the padding */
-                for (i=0;i<padding;i++){
-                    *ptr++=0;
-                    bytes--;
+                for (i = 0; i < padding; i++) {
+                    ptr[0] = 0;
+                    ptr += 1;
+                    bytes -= 1;
                 }
                 /* reset for the next color plane. */
                 if (state->x < planes * bytes_per_line) {
                     state->count = 1;
                     state->LAST = state->buffer[state->x];
-                    state->x++;
+                    state->x += 1;
                 }
-            }
+            } while (state->x < planes * bytes_per_line);
+
             /* read next line */
             state->state = FETCH;
             break;

libImaging/QuantOctree.c

@@ -407,7 +407,7 @@ int quantize_octree(Pixel *pixelData,
    /* remove the used fine colors from the coarse cube */
    subtract_color_buckets(coarseCube, paletteBucketsFine, nFineColors);
 
-   /* did the substraction cleared one or more coarse bucket? */
+   /* did the subtraction cleared one or more coarse bucket? */
    while (nCoarseColors > count_used_color_buckets(coarseCube)) {
       /* then we can use the free buckets for fine colors */
       nAlreadySubtracted = nFineColors;

libImaging/Resample.c

@@ -80,18 +80,52 @@ static struct filter LANCZOS = { lanczos_filter, 3.0 };
 #define PRECISION_BITS (32 - 8 - 2)
 
 
+UINT8 _lookups[512] = {
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
+    16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
+    32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
+    48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63,
+    64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79,
+    80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95,
+    96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111,
+    112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127,
+    128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143,
+    144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159,
+    160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175,
+    176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191,
+    192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207,
+    208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223,
+    224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239,
+    240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255
+};
+
+UINT8 *lookups = &_lookups[128];
+
+
 static inline UINT8 clip8(int in)
 {
-    if (in >= (1 << PRECISION_BITS << 8))
-       return 255;
-    if (in <= 0)
-        return 0;
-    return (UINT8) (in >> PRECISION_BITS);
+    return lookups[in >> PRECISION_BITS];
 }
 
 
 int
-ImagingPrecompute(int inSize, int outSize, struct filter *filterp,
+precompute_coeffs(int inSize, int outSize, struct filter *filterp,
                   int **xboundsp, double **kkp) {
     double support, scale, filterscale;
     double center, ww, ss;
@@ -142,6 +176,16 @@ ImagingPrecompute(int inSize, int outSize, struct filter *filterp,
         k = &kk[xx * kmax];
         for (x = 0; x < xmax; x++) {
             double w = filterp->filter((x + xmin - center + 0.5) * ss);
+            if (w == 0) {
+                if (x == 0) {
+                    x -= 1;
+                    xmin += 1;
+                    xmax -= 1;
+                } else if (x == xmax - 1) {
+                    xmax -= 1;
+                }
+                continue;
+            }
             k[x] = w;
             ww += w;
         }
@@ -162,6 +206,32 @@ ImagingPrecompute(int inSize, int outSize, struct filter *filterp,
 }
 
 
+int
+normalize_coeffs_8bpc(int outSize, int kmax, double *prekk, INT32 **kkp)
+{
+    int x;
+    INT32 *kk;
+
+    /* malloc check ok, overflow checked in precompute_coeffs */
+    kk = malloc(outSize * kmax * sizeof(INT32));
+    if ( ! kk) {
+        return 0;
+    }
+
+    for (x = 0; x < outSize * kmax; x++) {
+        if (prekk[x] < 0) {
+            kk[x] = (int) (-0.5 + prekk[x] * (1 << PRECISION_BITS));
+        } else {
+            kk[x] = (int) (0.5 + prekk[x] * (1 << PRECISION_BITS));
+        }
+    }
+
+    *kkp = kk;
+    return kmax;
+}
+
+
+
 Imaging
 ImagingResampleHorizontal_8bpc(Imaging imIn, int xsize, struct filter *filterp)
 {
@@ -170,28 +240,21 @@ ImagingResampleHorizontal_8bpc(Imaging imIn, int xsize, struct filter *filterp)
     int ss0, ss1, ss2, ss3;
     int xx, yy, x, kmax, xmin, xmax;
     int *xbounds;
-    int *k, *kk;
+    INT32 *k, *kk;
     double *prekk;
 
-
-    kmax = ImagingPrecompute(imIn->xsize, xsize, filterp, &xbounds, &prekk);
+    kmax = precompute_coeffs(imIn->xsize, xsize, filterp, &xbounds, &prekk);
     if ( ! kmax) {
         return (Imaging) ImagingError_MemoryError();
     }
 
-    kk = malloc(xsize * kmax * sizeof(int));
-    if ( ! kk) {
+    kmax = normalize_coeffs_8bpc(xsize, kmax, prekk, &kk);
+    free(prekk);    
+    if ( ! kmax) {
         free(xbounds);
-        free(prekk);
         return (Imaging) ImagingError_MemoryError();
     }
 
-    for (x = 0; x < xsize * kmax; x++) {
-        kk[x] = (int) (0.5 + prekk[x] * (1 << PRECISION_BITS));
-    }
-
-    free(prekk);
-
     imOut = ImagingNew(imIn->mode, xsize, imIn->ysize);
     if ( ! imOut) {
         free(kk);
@@ -282,28 +345,21 @@ ImagingResampleVertical_8bpc(Imaging imIn, int ysize, struct filter *filterp)
     int ss0, ss1, ss2, ss3;
     int xx, yy, y, kmax, ymin, ymax;
     int *xbounds;
-    int *k, *kk;
+    INT32 *k, *kk;
     double *prekk;
 
-
-    kmax = ImagingPrecompute(imIn->ysize, ysize, filterp, &xbounds, &prekk);
+    kmax = precompute_coeffs(imIn->ysize, ysize, filterp, &xbounds, &prekk);
     if ( ! kmax) {
         return (Imaging) ImagingError_MemoryError();
     }
 
-    kk = malloc(ysize * kmax * sizeof(int));
-    if ( ! kk) {
+    kmax = normalize_coeffs_8bpc(ysize, kmax, prekk, &kk);
+    free(prekk);    
+    if ( ! kmax) {
         free(xbounds);
-        free(prekk);
         return (Imaging) ImagingError_MemoryError();
     }
 
-    for (y = 0; y < ysize * kmax; y++) {
-        kk[y] = (int) (0.5 + prekk[y] * (1 << PRECISION_BITS));
-    }
-
-    free(prekk);
-
     imOut = ImagingNew(imIn->mode, imIn->xsize, ysize);
     if ( ! imOut) {
         free(kk);
@@ -396,7 +452,7 @@ ImagingResampleHorizontal_32bpc(Imaging imIn, int xsize, struct filter *filterp)
     int *xbounds;
     double *k, *kk;
 
-    kmax = ImagingPrecompute(imIn->xsize, xsize, filterp, &xbounds, &kk);
+    kmax = precompute_coeffs(imIn->xsize, xsize, filterp, &xbounds, &kk);
     if ( ! kmax) {
         return (Imaging) ImagingError_MemoryError();
     }
@@ -456,7 +512,7 @@ ImagingResampleVertical_32bpc(Imaging imIn, int ysize, struct filter *filterp)
     int *xbounds;
     double *k, *kk;
 
-    kmax = ImagingPrecompute(imIn->ysize, ysize, filterp, &xbounds, &kk);
+    kmax = precompute_coeffs(imIn->ysize, ysize, filterp, &xbounds, &kk);
     if ( ! kmax) {
         return (Imaging) ImagingError_MemoryError();
     }

libImaging/Unpack.c

@@ -747,6 +747,30 @@ unpackRGBa(UINT8* out, const UINT8* in, int pixels)
     }
 }
 
+static void
+unpackBGRa(UINT8* out, const UINT8* in, int pixels)
+{
+    int i;
+    /* premultiplied BGRA */
+    for (i = 0; i < pixels; i++) {
+        int a = in[3];
+        if (!a)
+            out[R] = out[G] = out[B] = out[A] = 0;
+        else if (a == 255) {
+            out[R] = in[2];
+            out[G] = in[1];
+            out[B] = in[0];
+            out[A] = a;
+        } else {
+            out[R] = CLIP(in[2] * 255 / a);
+            out[G] = CLIP(in[1] * 255 / a);
+            out[B] = CLIP(in[0] * 255 / a);
+            out[A] = a;
+        }
+        out += 4; in += 4;
+    }
+}
+
 static void
 unpackRGBAI(UINT8* out, const UINT8* in, int pixels)
 {
@@ -1206,6 +1230,7 @@ static struct {
     {"RGBA",    "LA;16B",       32,     unpackRGBALA16B},
     {"RGBA",    "RGBA",         32,     copy4},
     {"RGBA",    "RGBa",         32,     unpackRGBa},
+    {"RGBA",    "BGRa",         32,     unpackBGRa},
     {"RGBA",    "RGBA;I",       32,     unpackRGBAI},
     {"RGBA",    "RGBA;L",       32,     unpackRGBAL},
     {"RGBA",    "RGBA;15",      16,     ImagingUnpackRGBA15},

winbuild/config.py

@@ -28,9 +28,9 @@ libs = {
         'dir': 'tiff-4.0.6',
     },
     'freetype': {
-        'url': 'http://download.savannah.gnu.org/releases/freetype/freetype-2.6.3.tar.gz',
-        'hash': 'md5:8b0c80b64042b7c39b9fa9debe6305c3',
-        'dir': 'freetype-2.6.3',
+        'url': 'http://download.savannah.gnu.org/releases/freetype/freetype-2.6.5.tar.gz',
+        'hash': 'md5:31b2276515d9ee1c7f37d9c9f4f3145a',
+        'dir': 'freetype-2.6.5',
     },
     'lcms': {
         'url': SF_MIRROR+'/project/lcms/lcms/2.7/lcms2-2.7.zip',