Ensure bytestring is bytes before checking for gzip header

[marcelometal]

Fixes a TypeError when checking if input is gzipped using bytestring.startswith(b"\x1f\x8b").

File ".../site-packages/cairosvg/parser.py", line 447, in __init__
    if bytestring.startswith(b"\x1f\x8b"):
TypeError: startswith first arg must be str or a tuple of str, not bytes

This patch ensures bytestring is properly encoded as bytes before gzip header detection.

[marcelometal]

Hi @liZe ,

First of all, thanks for maintaining this great library.

This used to work in 2.7.1 but now breaks downstream code, for example in projects like Thumbor.

Would appreciate a review when possible. Thanks!

[liZe]

Hi!

This is probably already fixed in 2.8.1, did you try the latest version?

[marcelometal]

Hi!

This is probably already fixed in 2.8.1, did you try the latest version?

Yes, I tested it with the latest version (2.8.1), and the issue still occurs.

[liZe]

Yes, I tested it with the latest version (2.8.1), and the issue still occurs.

Could you please share a code sample that shows your problem?

[marcelometal]

I was able to reproduce it with this small example

import sys

import cairosvg

print(f"Python version: {sys.version}")
print(f"CairoSVG version: {cairosvg.__version__}")

svg_str = (
    "<svg width='100' height='100'><rect width='100' height='100'/></svg>"
)

cairosvg.svg2png(bytestring=svg_str)

Output:

Python version: 3.11.9 (main, Apr 10 2024, 13:16:36) [GCC 13.2.0]
CairoSVG version: 2.8.1
Traceback (most recent call last):
  File "test.py", line 12, in <module>
    cairosvg.svg2png(bytestring=svg_str)
  File "/home/metal/.virtualenvs/thumbor/lib/python3.11/site-packages/cairosvg/__init__.py", line 54, in svg2png
    return surface.PNGSurface.convert(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/metal/.virtualenvs/thumbor/lib/python3.11/site-packages/cairosvg/surface.py", line 129, in convert
    tree = Tree(
           ^^^^^
  File "/home/metal/.virtualenvs/thumbor/lib/python3.11/site-packages/cairosvg/parser.py", line 390, in __init__
    if bytestring.startswith(b'\x1f\x8b'):
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: startswith first arg must be str or a tuple of str, not bytes

[liZe]

I was able to reproduce it with this small example

Thanks. The bytestring argument was only supposed to accept… bytes, and not Unicode strings 😄.

Let’s fix that to avoid breaking other libraries!

[marcelometal]

Thank you @liZe !

On my side, I’ll make sure to fix it so that it only passes bytes — not Unicode strings.

Thanks again!