Merge branch 'main' into iss_1293

MartinThoma · web-flow · commit c67988e739c2 · 2022-09-02T07:47:26.000+02:00
diff --git a/PyPDF2/_codecs/adobe_glyphs.py b/PyPDF2/_codecs/adobe_glyphs.py
@@ -13425,3 +13425,13 @@
     "/zukatakana": "\u30BA",
     "/zwarakay": "\u0659",
 }
+
+
+def _complete() -> None:
+    global adobe_glyphs
+    for i in range(256):
+        adobe_glyphs[f"/a{i}"] = chr(i)
+    adobe_glyphs["/.notdef"] = "□"
+
+
+_complete()
diff --git a/PyPDF2/_page.py b/PyPDF2/_page.py
@@ -952,11 +952,11 @@ def scale(self, sx: float, sy: float) -> None:
         :param float sy: The scaling factor on vertical axis.
         """
         self.add_transformation((sx, 0, 0, sy, 0, 0))
-        self.mediabox = self.mediabox.scale(sx, sy)
         self.cropbox = self.cropbox.scale(sx, sy)
         self.artbox = self.artbox.scale(sx, sy)
         self.bleedbox = self.bleedbox.scale(sx, sy)
         self.trimbox = self.trimbox.scale(sx, sy)
+        self.mediabox = self.mediabox.scale(sx, sy)
         if PG.VP in self:
             viewport = self[PG.VP]
             if isinstance(viewport, ArrayObject):
diff --git a/PyPDF2/generic/_base.py b/PyPDF2/generic/_base.py
@@ -233,13 +233,10 @@ def __new__(
         try:
             return decimal.Decimal.__new__(cls, str_(value), context)
         except Exception:
-            try:
-                return decimal.Decimal.__new__(cls, str(value))
-            except decimal.InvalidOperation:
-                # If this isn't a valid decimal (happens in malformed PDFs)
-                # fallback to 0
-                logger_warning(f"Invalid FloatObject {value}", __name__)
-                return decimal.Decimal.__new__(cls, "0")
+            # If this isn't a valid decimal (happens in malformed PDFs)
+            # fallback to 0
+            logger_warning(f"FloatObject ({value}) invalid; use 0.0 instead", __name__)
+            return decimal.Decimal.__new__(cls, "0.0")
 
     def __repr__(self) -> str:
         if self == self.to_integral():
@@ -271,8 +268,11 @@ class NumberObject(int, PdfObject):
     NumberPattern = re.compile(b"[^+-.0-9]")
 
     def __new__(cls, value: Any) -> "NumberObject":
-        val = int(value)
-        return int.__new__(cls, val)
+        try:
+            return int.__new__(cls, int(value))
+        except ValueError:
+            logger_warning(f"NumberObject({value}) invalid; use 0 instead", __name__)
+            return int.__new__(cls, 0)
 
     def as_numeric(self) -> int:
         return int(repr(self).encode("utf8"))
diff --git a/docs/user/cropping-and-transforming.md b/docs/user/cropping-and-transforming.md
@@ -147,7 +147,7 @@ reader = PdfReader("resources/side-by-side-subfig.pdf")
 page = reader.pages[0]
 
 # Scale
-page.scale(0.5)
+page.scale_by(0.5)
 
 # Write the result to a file
 writer = PdfWriter()
diff --git a/tests/test_cmap.py b/tests/test_cmap.py
@@ -55,3 +55,11 @@ def test_bfchar_on_2_chars():
     reader = PdfReader(BytesIO(get_pdf_from_url(url, name=name)))
     for page in reader.pages:
         page.extract_text()
+
+
+def test_ascii_charset():
+    # iss #1312
+    url = "https://github.com/py-pdf/PyPDF2/files/9472500/main.pdf"
+    name = "ascii charset.pdf"
+    reader = PdfReader(BytesIO(get_pdf_from_url(url, name=name)))
+    assert "/a" not in reader.pages[0].extract_text()
diff --git a/tests/test_generic.py b/tests/test_generic.py
@@ -39,8 +39,14 @@
 RESOURCE_ROOT = PROJECT_ROOT / "resources"
 
 
-def test_float_object_exception():
+def test_float_object_exception(caplog):
     assert FloatObject("abc") == 0
+    assert caplog.text != ""
+
+
+def test_number_object_exception(caplog):
+    assert NumberObject("0,0") == 0
+    assert caplog.text != ""
 
 
 def test_number_object_no_exception():
