A DEFLATE-based stream compression/decompression library
This library provides support for compression and decompression of DEFLATE-based streams:
- the DEFLATE format itself
- the zlib format
These three formats are all closely related and largely only differ in their headers/footers. This crate has three types in each submodule for dealing with these three formats.
In addition to supporting three formats, this crate supports several different backends, controlled through this crate’s features:
rust_backend- this implementation uses the
miniz_oxidecrate which is a port of
miniz.c(below) to Rust. This feature does not require a C compiler and only requires Rust code.
zlib- this feature will enable linking against the
libzlibrary, typically found on most Linux systems by default. If the library isn’t found to already be on the system it will be compiled from source (this is a C library).
There’s various tradeoffs associated with each implementation, but in general you probably
won’t have to tweak the defaults. The default choice is selected to avoid the need for a C
compiler at build time.
zlib-ng-compat is useful if you’re using zlib for compatibility but
want performance via zlib-ng’s zlib-compat mode.
zlib is useful if something else in your
dependencies links the original zlib so you cannot use zlib-ng-compat. The compression ratios
and performance of each of these feature should be roughly comparable, but you’ll likely want
to run your own tests if you’re curious about the performance.
All types in the
write module work on instances of
whereas all types in the
read module work on instances of
bufread works with
BufRead. If you
are decoding directly from a
&[u8], use the
use flate2::write::GzEncoder; use flate2::Compression; use std::io; use std::io::prelude::*; let mut encoder = GzEncoder::new(Vec::new(), Compression::default()); encoder.write_all(b"Example")?;
Other various types are provided at the top-level of the crate for
management and dealing with encoders/decoders. Also note that types which
operate over a specific trait often implement the mirroring trait as well.
For example a
flate2::read::DeflateDecoder<T> also implements the
Write trait if
T: Write. That is, the “dual trait” is forwarded directly
to the underlying object if available.