Improved handling of PPM header

Tests/test_file_ppm.py

@@ -1,6 +1,6 @@
 import pytest
 
-from PIL import Image
+from PIL import Image, UnidentifiedImageError
 
 from .helper import assert_image_equal, assert_image_similar, hopper
 
@@ -50,6 +50,55 @@ def test_pnm(tmp_path):
             assert_image_equal(im, reloaded)
 
 
+def test_not_ppm(tmp_path):
+    path = str(tmp_path / "temp.ppm")
+    with open(path, "wb") as f:
+        f.write(b"PyInvalid")
+
+    with pytest.raises(UnidentifiedImageError):
+        with Image.open(path):
+            pass
+
+
+def test_header_with_comments(tmp_path):
+    path = str(tmp_path / "temp.ppm")
+    with open(path, "wb") as f:
+        f.write(b"P6 #comment\n#comment\r12#comment\r8\n128 #comment\n255\n")
+
+    with Image.open(path) as im:
+        assert im.size == (128, 128)
+
+
+def test_nondecimal_header(tmp_path):
+    path = str(tmp_path / "temp.ppm")
+    with open(path, "wb") as f:
+        f.write(b"P6\n128\x00")
+
+    with pytest.raises(ValueError):
+        with Image.open(path):
+            pass
+
+
+def test_token_too_long(tmp_path):
+    path = str(tmp_path / "temp.ppm")
+    with open(path, "wb") as f:
+        f.write(b"P6\n 0123456789")
+
+    with pytest.raises(ValueError):
+        with Image.open(path):
+            pass
+
+
+def test_too_many_colors(tmp_path):
+    path = str(tmp_path / "temp.ppm")
+    with open(path, "wb") as f:
+        f.write(b"P6\n1 1\n1000\n")
+
+    with pytest.raises(ValueError):
+        with Image.open(path):
+            pass
+
+
 def test_truncated_file(tmp_path):
     path = str(tmp_path / "temp.pgm")
     with open(path, "w") as f:

src/PIL/PpmImagePlugin.py

@@ -20,7 +20,7 @@
 #
 # --------------------------------------------------------------------
 
-b_whitespace = b"\x20\x09\x0a\x0b\x0c\x0d"
+B_WHITESPACE = b"\x20\x09\x0a\x0b\x0c\x0d"
 
 MODES = {
     # standard
@@ -49,26 +49,52 @@ class PpmImageFile(ImageFile.ImageFile):
     format = "PPM"
     format_description = "Pbmplus image"
 
-    def _token(self, s=b""):
+    def _read_magic(self, magic=b""):
         while True:  # read until next whitespace
             c = self.fp.read(1)
-            if not c or c in b_whitespace:
+            if c in B_WHITESPACE:
                 break
-            if c > b"\x79":
-                raise ValueError("Expected ASCII value, found binary")
-            s = s + c
-            if len(s) > 9:
-                raise ValueError("Expected int, got > 9 digits")
-        return s
+            magic += c
+            if len(magic) > 6:  # exceeded max magic number length
+                break
+        return magic
 
-    def _open(self):
+    def _read_token(self, token=b""):
+        def _ignore_comment():  # ignores rest of the line; stops at CR, LF or EOF
+            while self.fp.read(1) not in b"\r\n":
+                pass
+
+        while True:  # read until non-whitespace is found
+            c = self.fp.read(1)
+            if c == b"#":  # found comment, ignore it
+                _ignore_comment()
+                continue
+            if c in B_WHITESPACE:  # found whitespace, ignore it
+                if c == b"":  # reached EOF
+                    raise ValueError("Reached EOF while reading header")
+                continue
+            break
 
-        # check magic
-        s = self.fp.read(1)
-        if s != b"P":
-            raise SyntaxError("not a PPM file")
-        magic_number = self._token(s)
-        mode = MODES[magic_number]
+        token += c
+
+        while True:  # read until next whitespace
+            c = self.fp.read(1)
+            if c == b"#":
+                _ignore_comment()
+                continue
+            if c in B_WHITESPACE:  # token ended
+                break
+            token += c
+            if len(token) > 10:
+                raise ValueError(f"Token too long in file header: {token}")
+        return token
+
+    def _open(self):
+        magic_number = self._read_magic()
+        try:
+            mode = MODES[magic_number]
+        except KeyError:
+            raise SyntaxError("Not a PPM image file") from None
 
         self.custom_mimetype = {
             b"P4": "image/x-portable-bitmap",
@@ -83,29 +109,25 @@ def _open(self):
             self.mode = rawmode = mode
 
         for ix in range(3):
-            while True:
-                while True:
-                    s = self.fp.read(1)
-                    if s not in b_whitespace:
-                        break
-                    if s == b"":
-                        raise ValueError("File does not extend beyond magic number")
-                if s != b"#":
-                    break
-                s = self.fp.readline()
-            s = int(self._token(s))
-            if ix == 0:
-                xsize = s
-            elif ix == 1:
-                ysize = s
+            token = self._read_token()
+            try:  # check token sanity
+                token = int(token)
+            except ValueError:
+                raise ValueError(
+                    f"Non-decimal-ASCII found in header: {token}"
+                ) from None
+            if ix == 0:  # token is the x size
+                xsize = token
+            elif ix == 1:  # token is the y size
+                ysize = token
                 if mode == "1":
                     break
-            elif ix == 2:
-                # maxgrey
-                if s > 255:
+            elif ix == 2:  # token is maxval
+                maxval = token
+                if maxval > 255:
                     if not mode == "L":
-                        raise ValueError(f"Too many colors for band: {s}")
-                    if s < 2 ** 16:
+                        raise ValueError(f"Too many colors for band: {token}")
+                    if token < 2 ** 16:
                         self.mode = "I"
                         rawmode = "I;16B"
                     else: