zstd
zstd copied to clipboard
facebook
Zstandard - Fast real-time compression algorithm
Hi, not sure if it's really a bug though... but when I compile latest code on macOS12 targeting macOS10.15, the linker is not happy about huf_decompress_amd64.S.o here's my cmake line:...
**Describe the bug** This is a build issue that has come up after upgrading Visual Studio to v16.11.9. The build still worked with LLVM 13.0.0 and MSVC v16.11.7, but fails...
Not ready for merge yet; this has been tested across samples from silesia and canterbury. There are small gains in compression ratio as expected, however there are also losses, so...
**Is your feature request related to a problem? Please describe.** I'm experimenting with the "gzip" symlink on zstd binary and notice that there these compatibility options missing: -n --no-name When...
**Describe the bug** I wrote a simple test to produce a file with 1 million 4-byte incrementing ints, and tried to compress with various settings. I tried big and little...
This is a follow up to #2974, which was limited to `ZSTD_stableInBuffer` mode. In this PR, input data is just buffered, until a full block is reached, or a `flush()`...
This diff disables `tag` initialization when using the `rowHash` mode. This initialization was unconditional, but it becomes the dominating operation when compressing small data in streaming mode (see issue #2966)....
This PR bundles 3 modifications for high compression modes, targeting higher compression ratio on "small" (one-block) data. The benefit may extend beyond the first block, but they are vanishing (and...
**Describe the bug** I had expected no options would beat `--ultra -22`, however, I have found a file where it is outperformed by `pzstd -1 | pzstd -19`. **To Reproduce**...
Many of our Travis CI tests run on special hardware (like power pc) and cannot be migrated to GitHub actions (except possibly via QEMU, which isn't exactly equivalent). However, we...
