Bug report
Bug description:
We had a VM where ntp went very wonky and ended up thinking it was the year 2141!
Trying to write a gzip file with the system clock greater than 2106-02-07T06:28:15 results in a struct error because we tried to cram a 64-bit int into a 32 bit field (which won't work).
I wouldn't expect the gzip module to explode in this case, I'd expect it to set the MTIME to 0.
RFC 1952 (if that's at all relevant these days) states
MTIME = 0 means no time stamp is available
MWE
Only tested this on macOS on 3.13.2, but from a cursory glance, the mtime handling hasn't changed in over a decade.
>>> import gzip
>>> gzip.GzipFile("/dev/null", "w", mtime=2**32+1)
Traceback (most recent call last):
File "<python-input-3>", line 1, in <module>
gzip.GzipFile("/dev/null", "w", mtime=2**32+1)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.13/gzip.py", line 237, in __init__
self._write_gzip_header(compresslevel)
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.13/gzip.py", line 281, in _write_gzip_header
write32u(self.fileobj, int(mtime))
~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.13/gzip.py", line 77, in write32u
output.write(struct.pack("<L", value))
~~~~~~~~~~~^^^^^^^^^^^^^
struct.error: 'L' format requires 0 <= number <= 4294967295
Proposed Patch
I'd naïvely fix it with this patch
diff --git a/Lib/gzip.py b/Lib/gzip.py
index c00f51858de..29345de4659 100644
--- a/Lib/gzip.py
+++ b/Lib/gzip.py
@@ -297,6 +297,8 @@ def _write_gzip_header(self, compresslevel):
mtime = self._write_mtime
if mtime is None:
mtime = time.time()
+ if mtime > 4294967295:
+ mtime = 0
write32u(self.fileobj, int(mtime))
if compresslevel == _COMPRESS_LEVEL_BEST:
xfl = b'\002'
CPython versions tested on:
3.13
Operating systems tested on:
macOS
Linked PRs
Bug report
Bug description:
We had a VM where ntp went very wonky and ended up thinking it was the year 2141!
Trying to write a gzip file with the system clock greater than
2106-02-07T06:28:15results in a struct error because we tried to cram a 64-bit int into a 32 bit field (which won't work).I wouldn't expect the gzip module to explode in this case, I'd expect it to set the
MTIMEto 0.RFC 1952 (if that's at all relevant these days) states
MWE
Only tested this on macOS on 3.13.2, but from a cursory glance, the mtime handling hasn't changed in over a decade.
Proposed Patch
I'd naïvely fix it with this patch
CPython versions tested on:
3.13
Operating systems tested on:
macOS
Linked PRs