Failed to read JPX header

[myhloli]

Description of the bug

when i use the api page.get_image_rects，there is a RuntimeError: Failed to read JPX header

How to reproduce the bug

JFMA_15_11.pdf
page_id:75（page_id start from 0）

pdf_docs = fitz.open("pdf", pdf_bytes)
for page_id, page in enumerate(pdf_docs):
    page_imgs = page.get_images()
    for img in page_imgs:
    recs = page.get_image_rects(img, transform=True)

my code like this demo

recs = page.get_image_rects(img, transform=True)
           │    │               └ (1555, 0, 300, 28, 1, 'Indexed', '', 'Im1', 'JPXDecode')
           │    └ <function get_image_rects at 0x000001B37F685E40>
           └ page 75 of <memory, doc# 1>
pix = Pixmap(page.parent, xref)  # make pixmap of the image to compute MD5
          │      │    │       └ 1555
          │      │    └ <weakproxy at 0x000001B301178630 to Document at 0x000001B37E8D5910>
          │      └ page 75 of <memory, doc# 1>
          └ <class 'fitz.fitz.Pixmap'>
_fitz.Pixmap_swiginit(self, _fitz.new_Pixmap(*args))
    │     │               │     │     │           └ (<weakproxy at 0x000001B301178630 to Document at 0x000001B37E8D5910>, 1555)
    │     │               │     │     └ <built-in function new_Pixmap>
    │     │               │     └ <module 'fitz._fitz' from 
    │     │               └ <unprintable Pixmap object>
    │     └ <built-in function Pixmap_swiginit>
    └ <module 'fitz._fitz' from '

there is stack when i meet this error

PyMuPDF version

1.23.7

Operating system

Windows

Python version

3.11

[julian-smith-artifex-com]

Thanks for the reproducer.

It looks like the image with page=75 (zero-based) image=3 (zero-based) and xref=1555, is corrupted. But MuPDF isn't coping with particularly well with this - we end up getting this image's error for all images on page 75.

I'll mark this as an upstream bug and keep it open here.

[julian-smith-artifex-com]

Note that so far i've been testing this with PyMuPDF built with MuPDF master. It looks like MuPDF master may have a regression where a single corrupt JPX image causes PyMuPDF to return an error for all images on the same page.

But the current release of PyMuPDF-1.23.8 (which is built with MuPDF-1.23.7) appears to be handling things ok. The image with xref=1555 is corrupt, and it's returning an error for just that image. I think this is correct behaviour.

So i think this is not actually a bug in the current release after all.

[julian-smith-artifex-com]

Given that the image in question is corrupt, i think this isn't a bug after all, so i'll close it sometime soon.

Please comment here if you think there's still something wrong with pymupdf's behaviour.