From 7f8311efeaebe04d67b92757cc5eaae890b1794f Mon Sep 17 00:00:00 2001
From: Chris Hennick <hennickc@amazon.com>
Date: Thu, 11 Apr 2024 13:03:57 -0700
Subject: [PATCH] Add support for decompressing LZMA
---
Cargo.toml | 4 +++-
src/compression.rs | 10 ++++++++++
src/read.rs | 19 +++++++++++++++++++
src/read/lzma.rs | 47 ++++++++++++++++++++++++++++++++++++++++++++++
src/write.rs | 4 +++-
5 files changed, 82 insertions(+), 2 deletions(-)
create mode 100644 src/read/lzma.rs
diff --git a/Cargo.toml b/Cargo.toml
index 472daf36..7e8ecf36 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -28,6 +28,7 @@ time = { version = "0.3.34", optional = true, default-features = false, features
zstd = { version = "0.13.1", optional = true, default-features = false }
zopfli = { version = "0.8.0", optional = true }
deflate64 = { version = "0.1.8", optional = true }
+lzma-rs = { version = "0.3.0", optional = true }
[target.'cfg(any(all(target_arch = "arm", target_pointer_width = "32"), target_arch = "mips", target_arch = "powerpc"))'.dependencies]
crossbeam-utils = "0.8.19"
@@ -49,8 +50,9 @@ deflate-miniz = ["flate2/default"]
deflate-zlib = ["flate2/zlib"]
deflate-zlib-ng = ["flate2/zlib-ng"]
deflate-zopfli = ["zopfli"]
+lzma = ["lzma-rs/stream"]
unreserved = []
-default = ["aes-crypto", "bzip2", "deflate", "deflate64", "deflate-zlib-ng", "deflate-zopfli", "time", "zstd"]
+default = ["aes-crypto", "bzip2", "deflate", "deflate64", "deflate-zlib-ng", "deflate-zopfli", "lzma", "time", "zstd"]
[[bench]]
name = "read_entry"
diff --git a/src/compression.rs b/src/compression.rs
index 5352c487..14be540e 100644
--- a/src/compression.rs
+++ b/src/compression.rs
@@ -41,6 +41,9 @@ pub enum CompressionMethod {
/// Compress the file using ZStandard
#[cfg(feature = "zstd")]
Zstd,
+ /// Compress the file using LZMA
+ #[cfg(feature = "lzma")]
+ Lzma,
/// Unsupported compression method
#[cfg_attr(
not(fuzzing),
@@ -83,7 +86,10 @@ impl CompressionMethod {
pub const BZIP2: Self = CompressionMethod::Bzip2;
#[cfg(not(feature = "bzip2"))]
pub const BZIP2: Self = CompressionMethod::Unsupported(12);
+ #[cfg(not(feature = "lzma"))]
pub const LZMA: Self = CompressionMethod::Unsupported(14);
+ #[cfg(feature = "lzma")]
+ pub const LZMA: Self = CompressionMethod::Lzma;
pub const IBM_ZOS_CMPSC: Self = CompressionMethod::Unsupported(16);
pub const IBM_TERSE: Self = CompressionMethod::Unsupported(18);
pub const ZSTD_DEPRECATED: Self = CompressionMethod::Unsupported(20);
@@ -123,6 +129,8 @@ impl CompressionMethod {
9 => CompressionMethod::Deflate64,
#[cfg(feature = "bzip2")]
12 => CompressionMethod::Bzip2,
+ #[cfg(feature = "lzma")]
+ 14 => CompressionMethod::Lzma,
#[cfg(feature = "zstd")]
93 => CompressionMethod::Zstd,
#[cfg(feature = "aes-crypto")]
@@ -157,6 +165,8 @@ impl CompressionMethod {
CompressionMethod::Aes => 99,
#[cfg(feature = "zstd")]
CompressionMethod::Zstd => 93,
+ #[cfg(feature = "lzma")]
+ CompressionMethod::Lzma => 14,
CompressionMethod::Unsupported(v) => v,
}
diff --git a/src/read.rs b/src/read.rs
index c9291ab8..1eefca5a 100644
--- a/src/read.rs
+++ b/src/read.rs
@@ -37,6 +37,9 @@ use zstd::stream::read::Decoder as ZstdDecoder;
/// Provides high level API for reading from a stream.
pub(crate) mod stream;
+#[cfg(feature = "lzma")]
+pub(crate) mod lzma;
+
// Put the struct declaration in a private module to convince rustdoc to display ZipArchive nicely
pub(crate) mod zip_archive {
use std::sync::Arc;
@@ -81,6 +84,8 @@ pub(crate) mod zip_archive {
use crate::result::ZipError::InvalidPassword;
pub use zip_archive::ZipArchive;
+#[cfg(feature = "lzma")]
+use crate::read::lzma::LzmaReader;
#[allow(clippy::large_enum_variant)]
pub(crate) enum CryptoReader<'a> {
@@ -147,6 +152,8 @@ pub(crate) enum ZipFileReader<'a> {
Bzip2(Crc32Reader<BzDecoder<CryptoReader<'a>>>),
#[cfg(feature = "zstd")]
Zstd(Crc32Reader<ZstdDecoder<'a, io::BufReader<CryptoReader<'a>>>>),
+ #[cfg(feature = "lzma")]
+ Lzma(Crc32Reader<LzmaReader<CryptoReader<'a>>>),
}
impl<'a> Read for ZipFileReader<'a> {
@@ -168,6 +175,8 @@ impl<'a> Read for ZipFileReader<'a> {
ZipFileReader::Bzip2(r) => r.read(buf),
#[cfg(feature = "zstd")]
ZipFileReader::Zstd(r) => r.read(buf),
+ #[cfg(feature = "lzma")]
+ ZipFileReader::Lzma(r) => r.read(buf)
}
}
}
@@ -192,6 +201,11 @@ impl<'a> ZipFileReader<'a> {
ZipFileReader::Bzip2(r) => r.into_inner().into_inner().into_inner(),
#[cfg(feature = "zstd")]
ZipFileReader::Zstd(r) => r.into_inner().finish().into_inner().into_inner(),
+ #[cfg(feature = "lzma")]
+ ZipFileReader::Lzma(r) => {
+ let inner: Box<_> = r.into_inner().finish().unwrap().into();
+ Read::take(Box::leak(inner), u64::MAX)
+ }
}
}
}
@@ -313,6 +327,11 @@ pub(crate) fn make_reader(
let zstd_reader = ZstdDecoder::new(reader).unwrap();
ZipFileReader::Zstd(Crc32Reader::new(zstd_reader, crc32, ae2_encrypted))
}
+ #[cfg(feature = "lzma")]
+ CompressionMethod::Lzma => {
+ let reader = LzmaReader::new(reader);
+ ZipFileReader::Lzma(Crc32Reader::new(reader, crc32, ae2_encrypted))
+ }
_ => panic!("Compression method not supported"),
}
}
diff --git a/src/read/lzma.rs b/src/read/lzma.rs
new file mode 100644
index 00000000..5a7c41ac
--- /dev/null
+++ b/src/read/lzma.rs
@@ -0,0 +1,47 @@
+use std::collections::VecDeque;
+use std::io::{copy, Error, Read, Result, Write};
+use lzma_rs::decompress::{Options, Stream, UnpackedSize};
+
+const COMPRESSED_BYTES_TO_BUFFER: usize = 4096;
+
+const OPTIONS: Options = Options {
+ unpacked_size: UnpackedSize::ReadFromHeader,
+ memlimit: None,
+ allow_incomplete: true,
+};
+
+#[derive(Debug)]
+pub struct LzmaReader<R> {
+ compressed_reader: R,
+ stream: Stream<VecDeque<u8>>
+}
+
+impl <R: Read> LzmaReader<R> {
+ pub fn new(inner: R) -> Self {
+ LzmaReader {
+ compressed_reader: inner,
+ stream: Stream::new_with_options(&OPTIONS, VecDeque::new())
+ }
+ }
+
+ pub fn finish(mut self) -> Result<VecDeque<u8>> {
+ copy(&mut self.compressed_reader, &mut self.stream)?;
+ self.stream.finish().map_err(Error::from)
+ }
+}
+
+impl <R: Read> Read for LzmaReader<R> {
+ fn read(&mut self, buf: &mut [u8]) -> Result<usize> {
+ let mut bytes_read = self.stream.get_output_mut().unwrap().read(buf)?;
+ while bytes_read < buf.len() {
+ let mut next_compressed = [0u8; COMPRESSED_BYTES_TO_BUFFER];
+ let compressed_bytes_read = self.compressed_reader.read(&mut next_compressed)?;
+ if compressed_bytes_read == 0 {
+ break;
+ }
+ self.stream.write_all(&next_compressed[..compressed_bytes_read])?;
+ bytes_read += self.stream.get_output_mut().unwrap().read(&mut buf[bytes_read..])?;
+ }
+ Ok(bytes_read)
+ }
+}
\ No newline at end of file
diff --git a/src/write.rs b/src/write.rs
index 82305e4a..21d6f497 100644
--- a/src/write.rs
+++ b/src/write.rs
@@ -129,7 +129,7 @@ pub(crate) mod zip_writer {
pub(super) flush_on_finish_file: bool,
}
}
-use crate::result::ZipError::InvalidArchive;
+use crate::result::ZipError::{InvalidArchive, UnsupportedArchive};
use crate::write::GenericZipWriter::{Closed, Storer};
use crate::zipcrypto::ZipCryptoKeys;
use crate::CompressionMethod::Stored;
@@ -1269,6 +1269,8 @@ impl<W: Write + Seek> GenericZipWriter<W> {
GenericZipWriter::Zstd(ZstdEncoder::new(bare, level as i32).unwrap())
}))
}
+ #[cfg(feature = "lzma")]
+ CompressionMethod::Lzma => Err(UnsupportedArchive("LZMA isn't supported for compression")),
CompressionMethod::Unsupported(..) => {
Err(ZipError::UnsupportedArchive("Unsupported compression"))
}