page.get_text_block（）得到的结果无法解码

[chaodreaming]

from time import time
doc = fitz.open("pdf/base/base_5.pdf")
for i in range(len(doc)):
    page = doc.load_page(i)
    t1=time()
    for i in page.get_text_blocks():
        for j in i:
            print(j)

    print(time()-t1)``
![image](https://github.com/pymupdf/PyMuPDF/assets/49591435/dede3990-c20b-433f-a77a-2980d19b7c8c)
The result is a tuple, and I expected to get the formula characters inside. It cannot print the formula characters like a PDF miner. I have tried many ways, but I cannot obtain txt from the tuple

[chaodreaming]

information
[37.58697509765625, 319.2457275390625, 288.6783752441406, 406.44970703125, 'splits the ordinary convolution into two parts and utilizes fewer filters\nto generate several intrinsic feature maps. Then, a certain number of\ncheap transformation operations will be further applied for generating\nghost feature maps efficiently. During the implementation of the Ghost\nmodule, the required number of FLOPs can be calculated as 𝑖\udc5b∕𝑖\udc60⋅ℎ′ ⋅𝑖\udc64′ ⋅\n𝑎\udc50 ⋅ 𝑖\udc58 ⋅ 𝑖\udc58 + (𝑖\udc60 − 1) ⋅ n∕s ⋅ ℎ′ ⋅ 𝑖\udc64′ ⋅ 𝑎\udc51 ⋅ 𝑎\udc51, where 𝑎\udc51 × 𝑎\udc51 refers to the size of the\nkernel of linear operations (has the similar magnitude as that of 𝑖\udc58 × 𝑖\udc58),\n𝑖\udc60 refers to the number of cheap transformation operations, and 𝑖\udc60 ≪ 𝑎\udc50.\n', 2, 0]

[JorjMcKie]

for b in page.get_text("blocks"):
    print(b[4])

Looking at the documentation exhibits, that b[4] is the text.

[chaodreaming]

base_5.pdf
Traceback (most recent call last):
File "D:\pdf_layout\试试fitz获取位置信息.py", line 9, in
print(b[4])
UnicodeEncodeError: 'utf-8' codec can't encode character '\udc5b' in position 349: surrogates not allowed

[chaodreaming]

I have uploaded the file, and if only I have this issue, I will submit a bug

[JorjMcKie]

Please submit a bug to MuPDF's system: https://bugs.ghostscript.com/enter_bug.cgi.
Mention that the output of mutool -o test.txt base_5.pdf does not meet requirements.

[chaodreaming]

Why not the current warehouse

[JorjMcKie]

what does that mean?

[chaodreaming]

Why not submit an error to PyMuPDF

[JorjMcKie]

because the error is in the base library, not in PyMuPDF

[chaodreaming]

Forget it, I'll use another library