issue #2959: improve Info setting and dumping

Tests/test_file_pdf.py

@@ -137,13 +137,13 @@ class TestFilePdf(PillowTestCase):
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 1)
         self.assertEqual(len(pdf.info), 1)
-        self.assertEqual(pdfParser.decode_text(pdf.info[b"Producer"]), "pdfParser")
+        self.assertEqual(pdfParser.decode_text(pdf.info["Producer"]), "pdfParser")
         # append some info
-        pdf.info[b"Title"] = pdfParser.encode_text("abc")
-        pdf.info[b"Author"] = pdfParser.encode_text("def")
-        pdf.info[b"Subject"] = pdfParser.encode_text("ghi")
-        pdf.info[b"Keywords"] = pdfParser.encode_text("jkl")
-        pdf.info[b"Creator"] = pdfParser.encode_text("hopper()")
+        pdf.info["Title"] = "abc"
+        pdf.info["Author"] = "def"
+        pdf.info["Subject"] = "ghi"
+        pdf.info["Keywords"] = "jkl"
+        pdf.info["Creator"] = "hopper()"
         with open(pdf_filename, "r+b") as f:
             f.seek(0, os.SEEK_END)
             pdf.write_xref_and_trailer(f)
@@ -151,7 +151,7 @@ class TestFilePdf(PillowTestCase):
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 1)
         self.assertEqual(len(pdf.info), 6)
-        self.assertEqual(pdfParser.decode_text(pdf.info[b"Title"]), "abc")
+        self.assertEqual(pdfParser.decode_text(pdf.info["Title"]), "abc")
         # append two images
         mode_CMYK = hopper("CMYK")
         mode_P = hopper("P")
@@ -160,8 +160,8 @@ class TestFilePdf(PillowTestCase):
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 3)
         self.assertEqual(len(pdf.info), 6)
-        self.assertEqual(pdfParser.decode_text(pdf.info[b"Title"]), "abc")
-        self.assertEqual(pdfParser.decode_text(pdf.info[b"Producer"]), "pdfParser")
+        self.assertEqual(pdfParser.decode_text(pdf.info["Title"]), "abc")
+        self.assertEqual(pdfParser.decode_text(pdf.info["Producer"]), "pdfParser")
 
     def test_pdf_append_to_bytesio(self):
         im = hopper("RGB")

src/PIL/PdfImagePlugin.py

@@ -62,17 +62,17 @@ def _save(im, fp, filename, save_all=False):
         existing_pdf = pdfParser.PdfParser()
 
     if title:
-        existing_pdf.info[b"Title"] = pdfParser.encode_text(title)
+        existing_pdf.info["Title"] = title
     if author:
-        existing_pdf.info[b"Author"] = pdfParser.encode_text(author)
+        existing_pdf.info["Author"] = author
     if subject:
-        existing_pdf.info[b"Subject"] = pdfParser.encode_text(subject)
+        existing_pdf.info["Subject"] = subject
     if keywords:
-        existing_pdf.info[b"Keywords"] = pdfParser.encode_text(keywords)
+        existing_pdf.info["Keywords"] = keywords
     if creator:
-        existing_pdf.info[b"Creator"] = pdfParser.encode_text(creator)
+        existing_pdf.info["Creator"] = creator
     if producer:
-        existing_pdf.info[b"Producer"] = pdfParser.encode_text(producer)
+        existing_pdf.info["Producer"] = producer
 
     #
     # make sure image data is available

src/PIL/pdfParser.py

@@ -221,24 +221,10 @@ class PdfArray(list):
 
 
 class PdfDict(UserDict):
-    #def __init__(self, *args, orig_ref=None, pdf=None, **kwargs):
     def __init__(self, *args, **kwargs):
         UserDict.__init__(self, *args, **kwargs)
-        #self.orig_ref = kwargs.pop("orig_ref", None)
-        #self.pdf = kwargs.pop("pdf", None)
-        #self.is_changed = False
-
-    def __setitem__(self, key, value):
-        self.is_changed = True
-        UserDict.__setitem__(self, key, value)
 
     def __bytes__(self):
-        #if self.orig_ref is not None:
-        #    if self.is_changed:
-        #        if self.pdf is not None:
-        #            del self.pdf.xref_table[self.orig_ref.object_id]
-        #    else:
-        #        return bytes(self.orig_ref)
         out = bytearray(b"<<")
         for key, value in self.items():
             if value is None:
@@ -248,17 +234,12 @@ class PdfDict(UserDict):
             out.extend(bytes(PdfName(key)))
             out.extend(b" ")
             out.extend(value)
-            #out += b"\n%s %s" % (PdfName(key), value)
         out.extend(b"\n>>")
         return bytes(out)
         #return out + b"\n>>"
 
-    __str__ = __bytes__
-
-    #def write(self, f, orig_ref=None, pdf=None):
-    #    self.orig_ref = orig_ref
-    #    self.pdf = pdf
-    #    f.write(bytes(self))
+    if str == bytes:
+        __str__ = __bytes__
 
 
 class PdfBinary:
@@ -290,7 +271,7 @@ def pdf_repr(x):
     elif isinstance(x, list):
         return bytes(PdfArray(x))
     elif isinstance(x, str) and str != bytes:
-        return pdf_repr(x.encode("utf-8"))
+        return pdf_repr(encode_text(x))
     elif isinstance(x, bytes):
         return b"(" + x.replace(b"\\", b"\\\\").replace(b"(", b"\\(").replace(b")", b"\\)") + b")"  # XXX escape more chars? handle binary garbage
     else: