New feature created: .7zignore
Hello, I created a new archive from the 7-zip app and added an mp3 while selecting 7-zip password-protection on compression. Then I dragged and dropped other mp3s into the archive, while open in the app, from other windows folders. Afterwards, if I go into the archive in file explorer, the first mp3 does not open because it is encrypted, but all the other mp3s do open. If I do the exact same thing with pdfs, all of the pdfs will be encrypted. It therefore seems to be a problem with mp3 files. I am...
7zip installpath parameter is not used but visible in install log
CVE-2026-48102
CVE-2026-48101
Feature ".7zignore" 100% functional
7z password open
GPU Support
I'll fix some of these things. But that 64-bit / 32-bit difference can be ignored by client code. There is vt field inside PROPVARIANT that contains correct value for each returned item.
I'll fix some of these things. But that 64-bit / 32-bit difference can be ignored by client code. There is vt field inside PROPVARIANT that contaions correct value for each returned item.
looks like the same thing with kpidNumAltStreams aswell? VT_UI8 in k7z_PROPID_To_VARTYPE, assigned UInt32 in wim and ntfs handlers
kpidStreamId also looks to be inconsistent, its defined as VT_UI8 in k7z_PROPID_To_VARTYPE, ExtHandler.cpp, CpioHandler.cpp and WimHandler.cpp assign a UInt32, however NtfsHandler.cpp assigns a UInt64 instead to the property
sorry, that should say the declared VT type in k7z_PROPID_To_VARTYPE in PropID.cpp prop is defined as VT_UI8
kPidINode is reported as VT_UI8 but returned as VT_UI4 in archive handlers
Password from Environment Variable
Thanks for your reply! I've send a report to Samsung's CEO Office regarding the headers error and they said they will foward it to relevant departement.
If there are errors in zip64 headers, you can write to samsung and ask them to fix these errors.
Headers Error Warning when Extracting Samsung Firmware
If you're viewing directly on the SourceForge site, please note that the markdown preview is cutting off parts of the image - you'll need to open it in a new tab to see it in its whole.
Support raw image files with MBR records that only set LBA values
Speed benefits of choosing physical cores over SMT/HyperThreading logical cores
Sequence of items: struct Item { char Magic[8]; // "KSIZIP" uint32 Unknown; // Probably some flags uint32 Size; uint32 PackedSize; FILETIME FT; // Probably the modification date string Name; // Null-terminated string byte Data[PackedSize]; // ZLIB stream };
Sequence of items: struct Item { char Magic[8]; // "KSIZIP" uint32 Unknown; // Probably some flags uint32 Size; uint32 PackedSize; FILETIME FT; // Probably the modification date string Name; byte Data[Size]; // ZLIB stream };
Sequence of items: struct Item { char Magic[8]; uint32 Unknown; // Probably some flags uint32 Size; uint32 PackedSize; FILETIME FT; // Probably the modification date string Name; byte Data[Size]; // ZLIB stream };
Sequence of items: struct Item { char Magic[8]; uint32 Unknown; // Probably some flags uint32 Size; uint32 PackedSize; FILETIME FT; // Probably the modification date string Name; // byte Data[Size]; // ZLIB stream };
Thanks for creating this, I really appreciate it. Out of curiosity, what is the general structure of the format?
The archive format is very simple. I created a small plugin for opening these packages: https://www.tc4shell.com/en/7zip/ksizip7z/
For a bit of background, the BrailleNote was a specialized PDA for the blind that ran Windows CE. The device includes installers for various versions of the user guide among other things, and I thought it would be nice if 7-zip could extract files from these installers. I suspect the data is compressed as a zip file with a specialized header, however I am not entirely sure as my reverse engineering knowledge is limited. I attached an archive of some of these installers for possible investigation...
When I open or create a 7z archive I noticed that the "Packed Size" column is mostly blank. See attached. If I open an item in the archive and then update it, and then tell 7z to update the archive then that entry also then shows the Packed Size correctly. Is this a reportable bug? Thanks!
When I open or create a 7z archive I noticed that the "Packed Sizxe" column is mostly blank. See attached. If I open an item in the archive and then update it, and then tell 7z to update the archive then that entry also then shows the Packed Size correctly. Is this a reportable bug? Thanks!
Could you update your release notes pls, You allegedly fixed this CVE-2026-48095 :) https://securitylab.github.com/advisories/GHSL-2026-140_7-Zip/
Referencing OOB memory in LvmHandler.cpp
Referencing OOB memory in LvmHandler.cpp
Please disregard this post as it was posted prematurely!
Wrong Encryption Password and Errorlevel Value
Wrong Password and errorlevel value
You can select several files with Ctrl + Left Mouse Click
Support extracting first N bytes of files via stdout
That's not it. In this case, deleting each file will unpack and compress the archive. What if there are 20 files—unpack, delete, and repack them 20 times? It would be convenient to select several files with different paths in the file manager (for example, by creating a temporary file list) so you can unpack, delete, and repack them all in one go.
You can see full list of files in archive with menu command "View / Flat View" Then you can select any of them and delete.
Mark multiple files in an archive for deletion.
Found the problem. It looks like 7zip already preallocates the files correctly, and it works. The problem (excess fragmentation) only happens when the target is an NTFS-compressed folder, so it's NTFS issue, not 7zip issue. Sorry for the false report!
Found the problem. It looks like 7zip already preallocates the files correctly, and it works. The problem (excess fragmentation) only happens when when the target is an NTFS-compressed folder, so it's NTFS issue, not 7zip issue. Sorry for the false report!
show example
Incorrect info in 7zFM (26.01) about Differencing VHDs
check it again, and describe all steps and commands including name of archive. try same actions in another programs. use latest 7-Zip 26.01.
check it again, and describe all steps including name of archive. try same actions in another programs.
Writing 1 byte into a region of size 0 in Bcj2Enc.c
false positive
7za e a.zip -so > newname.txt
look sizes
Defraggler.
Writing 1 byte into a region of size 0 in Bcj2Enc.c
I'd guess they probably used a tool such as contig ( https://learn.microsoft.com/en-us/sysinternals/downloads/contig ) with the "-a" switch, or just happened to notice it while looking at a 3rd party defragmenter that showed it.
Files getting changed in the Zip file after origianally getting zipped
This is still a problem with the command line 26.01. I had to create an expect script to be able to supply a password file to 7zz on Linux to avoid exposing the password to ps (and to suppress it echoing the password).
I translated 7-Zip into Rusyn language and would like to contribute this translation to the official project.
How did you measure that extracted 80 MB file has 680 fragments?
Softsasi has recently built a brand new website for 7-Zip. Check it out! 🌐 Live Website: https://7zip.softsasi.com 💻 GitHub Repository: https://github.com/Softsasi/7zip This project includes many new improvements compared to the previous version, including: Enhanced UI/UX and modern animations Additional FAQ sections Community tab improvements with Issue Creator & Pull Request Creator information Better responsiveness and smoother interactions Cleaner structure and improved overall experience If...
Dynamic VHD parser: real vs supposed behavior
Could you add a CLI option to set the `removePathParts`?
This particular one (method: 04F71101) is Zstd \> 7z.exe i | grep 4F71101 0 ED 4F71101 ZSTD and was made with 7-Zip-zstd. One can still decompress it with it or just run it as it's self extracting exe. Maybe 7-zip should consider adding this method as supported into main line. After all it can decompress zstd archives. It could also decompress is as a codec.
The key derivation method is consistent with AES-256. The registry records the XChaCha20 option. Use a 64-bit counter to support very large files. @ipavlov https://github.com/ip7z/7zip/pull/215
It's annoying that every file extracted by 7zip on Windows (to an NTFS hard drive) is highly fragmented: 80 MB file = 680 fragments. To avoid this, the file should be pre-resized. I don't know the exact API, but I can look it up, it should be a very simple change (assuming you know the size of the file before writing to it).