https://bugs.kde.org/show_bug.cgi?id=403899
--- Comment #5 from Stefan Brüns <stefan.bru...@rwth-aachen.de> --- Git commit 2bf86d9d8e078513cc5d9f00d8e95544d9990d6c by Stefan Brüns. Committed on 31/03/2025 at 20:04. Pushed by bruns into branch 'master'. Handle Zip64 extra field in central directory Fix two problems with the current handling of size and relative header offset in the Central Directory, when these have the special "invalid" 0xffffffff marker value: - The "Local File Header" may also have "invalid" values for compressed/ uncompressed size, i.e. when the sizes are recorded in a succeeding Data Descriptor. Use the Zip64 Extra field, which must be present in this case. - The "Local Header Offset" in the "Central Directory" may be invalid as well, also use the Zip64 Extra field. The fields in the Central Directory MAY also use the "invalid" value even if case size and offset would fit in the 32 bit value, and always use the Zip64 format. Add two test files, the first exceeds 4 GByte uncompressed and requires Zip64 format: $> (echo -en abcd ; dd if=/dev/zero bs=1M count=4200 ; echo -en ABCD) \ | zip - - | gzip -c - > zip64_extra_zip64_size.zip.gz The second was created by Windows XPS virtual printer. It uses Zip64 fields although not necessary, all sizes/offsets are small. A +- -- autotests/data/zip64_extra_zip64_localheader.oxps A +- -- autotests/data/zip64_extra_zip64_size.zip.gz M +54 -0 autotests/karchivetest.cpp M +2 -0 autotests/karchivetest.h M +41 -44 src/kzip.cpp https://invent.kde.org/frameworks/karchive/-/commit/2bf86d9d8e078513cc5d9f00d8e95544d9990d6c -- You are receiving this mail because: You are watching all bug changes.
