Open nico opened 8 months ago
(This is responsible for all this diag in test_pdf.py 0000
:
Internal error while processing PDF file: Reached an unsupported marker, in 1 files, on 8 pages, 8 times
/Users/thakis/Downloads/0000/0000930.pdf 1 2 3 4 5 6 7 8
)
We decode it correctly (AFAICS) with: https://github.com/SerenityOS/serenity/pull/19571
And it seems to be the correct fix, as GIMP reports this error when I open the file:
Corrupt JPEG data: 4394 extraneous bytes before marker 0xd9
With some additional debug logging (https://github.com/nico/serenity/pull/new/jpeg-debug):
Looks like there's some trailing data (…or we read to little data) before the
JPEG_EOI
after we're done withdecode_huffman_stream()
.