severity 1095315 normal
thanks
On Fri, Feb 28, 2025 at 10:09:27AM +0000, Jelmer Vernooij wrote:
On Thu, Feb 06, 2025 at 03:37:51PM -0800, Otto Kekäläinen wrote:
Package: debmutate
Version: 0.71
Severity: important
Tags: python3.13
The python3-debmutate package fails to handle UTF-8 encoded files when
running under Python 3.13. This affects multiple Debian packaging
tools including lintian-brush, making it impossible to process
packages containing non-ASCII characters.
When running lintian-brush, the following error occurs:
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position
705: ordinal not in range(128)
I'm having trouble reproducing this:
$ ldd ~/src/lintian-brush/target/debug/lintian-brush
linux-vdso.so.1 (0x00007fde2f1fb000)
libpython3.13.so.1.0 => /lib/x86_64-linux-gnu/libpython3.13.so.1.0
(0x00007fde2da00000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1
(0x00007fde2e1a5000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fde2d91a000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fde2d724000)
/lib64/ld-linux-x86-64.so.2 (0x00007fde2f1fd000)
libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007fde2d704000)
libexpat.so.1 => /lib/x86_64-linux-gnu/libexpat.so.1
(0x00007fde2d6d9000)
$ ~/src/lintian-brush/target/debug/lintian-brush --no-update-changelog --modern
--uncertain
█████████████████████████████████████████████████████████████████████████████████████████
154/154No changes made.
By default Python uses utf-8 still, but with some environment
variables set it will change to ascii:
$ cat x.py
#!/usr/bin/python3
with open('debian/copyright', 'r') as f:
f.read()
$ LC_ALL=C LANG=C PYTHONUTF8=0 python3.12 x.py
Traceback (most recent call last):
File "/tmp/mariadb/x.py", line 4, in <module>
f.read()
File "/usr/lib/python3.12/encodings/ascii.py", line 26, in decode
return codecs.ascii_decode(input, self.errors)[0]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc2 in position 5873:
ordinal not in range(128)
We should handle this properly even if python is forced into ascii mode though.
Jelmer
