From fc7eea9a50f8e9c6dbced3de5478ba3fe4540148 Mon Sep 17 00:00:00 2001
From: David Huggins-Daines <dhd@ecolingui.ca>
Date: Thu, 28 Nov 2024 10:53:32 -0500
Subject: [PATCH] fix: require a newline before EI to fix various inline images
 (#25)

it might be good to check what other PDF readers do as the spec
is super unclear on this and unfortunately some creators do
make inline images with "EI" in the data and no filter...
---
 playa/parser.py             | 28 +++++++++++++++++-----------
 tests/test_object_parser.py |  2 +-
 2 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/playa/parser.py b/playa/parser.py
index 88da36f..0ff4c44 100644
--- a/playa/parser.py
+++ b/playa/parser.py
@@ -170,7 +170,7 @@ def nextline(self) -> Tuple[int, bytes]:
         return (linepos, self.data[linepos : self.pos])
 
     def get_inline_data(
-        self, target: bytes = b"EI", blocksize: int = -1
+        self, target: bytes = b"\nEI", blocksize: int = -1
     ) -> Tuple[int, bytes]:
         """Get the data for an inline image up to the target
         end-of-stream marker.
@@ -349,6 +349,7 @@ def __next__(self) -> StackEntry:
             if self.stack and top is None:
                 return self.stack.pop()
             (pos, token) = self.nexttoken()
+            log.debug("token at %d: %r", pos, token)
             if token is KEYWORD_ARRAY_BEGIN:
                 if top is None:
                     top = pos
@@ -429,7 +430,7 @@ def __next__(self) -> StackEntry:
                 dic = {
                     literal_name(k): v for (k, v) in choplist(2, objs) if v is not None
                 }
-                eos = b"EI"
+                eos = b"\nEI"
                 filter = dic.get("F")
                 if filter is not None:
                     if not isinstance(filter, list):
@@ -441,16 +442,19 @@ def __next__(self) -> StackEntry:
                 # operator shall be followed by a single white-space
                 # character, and the next character shall be
                 # interpreted as the first byte of image data.
-                if eos == b"EI":
+                if eos == b"\nEI":
                     self.seek(idpos + len(KEYWORD_ID.name) + 1)
                     (eipos, data) = self.get_inline_data(target=eos)
-                    # FIXME: it is totally unspecified what to do with
-                    # a newline between the end of the data and "EI",
-                    # since there is no explicit stream length.  (PDF
-                    # 1.7 p. 756: There should be an end-of-line
-                    # marker after the data and before endstream; this
-                    # marker shall not be included in the stream
-                    # length.)  We will include it, which might be wrong.
+                    log.debug("data at %d: %r", eipos, data)
+                    # It is totally unspecified what to do with a
+                    # newline between the end of the data and "EI",
+                    # since there is no explicit stream length, but in
+                    # practice, there is always a newline, since "EI"
+                    # may occur in the stream data even when it is not
+                    # ASCII85, so we will not include it.  (PDF 1.7
+                    # p. 756: There *should* be an end-of-line marker
+                    # after the data and before endstream; this marker
+                    # shall not be included in the stream length.)
                     data = data[: -len(eos)]
                 else:
                     # Note absence of + 1 here (the "Unless" above)
@@ -468,7 +472,9 @@ def __next__(self) -> StackEntry:
                 log.debug("InlineImage @ %d: %r", pos, obj)
                 # Inline images must occur at the top level, otherwise
                 # something is wrong (probably a corrupt file)
-                assert pos == top, f"Inline image {obj} not at top level of stream"
+                assert (
+                    pos == top
+                ), f"Inline image {obj} not at top level of stream ({pos} != {top}, {self.stack})"
                 top = None
                 return pos, obj
             else:
diff --git a/tests/test_object_parser.py b/tests/test_object_parser.py
index 23eea93..eda8513 100644
--- a/tests/test_object_parser.py
+++ b/tests/test_object_parser.py
@@ -394,7 +394,7 @@ def test_inline_images():
     pos, img = next(parser)
     assert isinstance(img, InlineImage)
     assert img.attrs["Foo"] == b"bar"
-    assert img.rawdata == b"VARIOUS UTTER NONSENSE\n"
+    assert img.rawdata == b"VARIOUS UTTER NONSENSE"
     pos, img = next(parser)
     assert isinstance(img, InlineImage)
     assert img.buffer == b"VARIOUS UTTER NONSENSE"