A malformed PDF may store a non-dictionary object (an array, name, or
other type) at /Resources or /Resources /XObject. The pdfinfo image
scanner iterated these with .items()/.as_dict() and probed them with the
`in` operator, which raise TypeError/ValueError on non-dictionary pikepdf
objects and crashed PdfInfo on otherwise-processable files. OCRmyPDF's
domain is messy, machine-generated PDFs, so scanning must tolerate this.
Guard _image_xobjects and _find_form_xobject_images with
isinstance(x, Dictionary) before iterating, treating a non-dictionary
/Resources or /XObject as "no image XObjects". This is the same
robustness class as the pdfa.py find_nonembedded_cid_fonts fix, applied
to the pdfinfo image scanner.
Address final review findings: a mask painted inside a Form XObject now
inherits the fill color in effect at the form's Do operator (previously it
reset to black, missing gray/color promotion one indirection deep). The cs
operator now resets the fill color to black per PDF spec, so a stale color
set before cs cannot leak to a subsequently drawn mask.
Track the current PDF fill color on the graphics stack alongside the CTM and
record an Ink classification (mono/gray/color) per image-draw event. Image
masks are painted with the current fill color, so this enables later device
promotion. Color operators are tolerant of malformed operands to preserve
robustness on untrusted input.
FlateDecode+DCTDecode compressed images are essentially deflated JPEGs,
typically created by OCRmyPDF's optimizer. This change ensures pdfinfo
correctly identifies them and should_visible_page_image_use_jpg treats
them as JPEG-origin images, allowing JPEG output when appropriate.
Split the 1288-line info.py into smaller, single-responsibility modules:
- _types.py: Enums, type aliases, lookup dictionaries
- _contentstream.py: PDF content stream parsing, DPI calculation
- _image.py: ImageInfo class and image finding functions
- _worker.py: Concurrency/worker process handling
- info.py: PageInfo, PdfInfo classes (reduced to ~530 lines)
Public API unchanged - all existing imports continue to work.
Full message
eportlab/lib/rl_safe_eval.py:11: DeprecationWarning: ast.NameConstant is deprecated and will be removed in Python 3.14; use ast.Constant instead
haveNameConstant = hasattr(ast,'NameConstant')
Warning triggered by reportlab-4.0.7 and Python 3.12
Our method of getting data from pdfminer would silently consume a StopIteration
if pdfminer returned no processed pages, leading to odd error message.
We improve an error from pdfminer properly, and returning a more
descriptive error of our own.
It would be possible for ocrmypdf to repair the file before sending it to
pdfminer, but this seems to be rare enough that we won't do that yet.
While faster than Python based methods, we've outgrown the limited
amount of information Ghostscript provides with this feature, and it
repeats an analysis we have to do anyway to learn what images are
present.
James R. Barlow
James Barlow
James R. Barlow