]>
code.octet-stream.net Git - hashgood/blob - src/calculate.rs
efc39860aa4a3fb3e9eeec9ac0bbc83f2711cd2a
2 use crossbeam_channel
::bounded
;
3 use crossbeam_channel
::Receiver
;
4 use crypto
::digest
::Digest
;
6 use crypto
::sha1
::Sha1
;
7 use crypto
::sha2
::Sha256
;
10 use std
::io
::prelude
::*;
14 use std
::thread
::JoinHandle
;
16 pub type CalculateResult
= Result
<Vec
<(Algorithm
, Vec
<u8>)>, Box
<dyn Error
>>;
18 /// For a given path to the input (may be "-" for STDIN), try to obtain a reader for the data within it.
19 pub fn get_input_reader(input
: &Path
) -> Result
<Box
<dyn Read
>, String
> {
20 if input
.to_str() == Some("-") {
21 // Special case: standard input
22 return Ok(Box
::new(std
::io
::stdin()));
26 "The path '{}' does not exist.",
27 input
.to_string_lossy()
32 "The path '{}' is not a regular file.",
33 input
.to_string_lossy()
36 match File
::open(input
) {
37 Ok(f
) => Ok(Box
::new(f
)),
38 Err(e
) => Err(format
!("File open: {}", e
)),
42 /// For the given input stream, calculate all requested digest types
43 pub fn create_digests(algorithms
: &[Algorithm
], mut input
: Box
<dyn Read
>) -> CalculateResult
{
44 let mut senders
= vec
![];
45 let mut handles
= vec
![];
47 if algorithms
.contains(&Algorithm
::Md5
) {
48 let (s
, r
) = bounded
::<Arc
<Vec
<u8>>>(1);
50 handles
.push(md5_digest(r
));
52 if algorithms
.contains(&Algorithm
::Sha1
) {
53 let (s
, r
) = bounded
::<Arc
<Vec
<u8>>>(1);
55 handles
.push(sha1_digest(r
));
57 if algorithms
.contains(&Algorithm
::Sha256
) {
58 let (s
, r
) = bounded
::<Arc
<Vec
<u8>>>(1);
60 handles
.push(sha256_digest(r
));
63 // 64 KB chunks will be read from the input at 64 KB and supplied to all hashing threads at once
64 // Right now that could be up to three threads. If CPU-bound, the other threads will mostly block while the slowest one finishes
65 const BUF_SIZE
: usize = 1024 * 64;
66 let mut buf
= [0; BUF_SIZE
];
67 while let Ok(size
) = input
.read(&mut buf
) {
71 // Create a shared read-only copy for the hashers to take as input
72 // buf is freed up for more reading
73 let chunk
= Arc
::new(buf
[0..size
].to_vec());
75 s
.send(chunk
.clone())?
;
80 // Once all data has been sent we just have to wait for the digests to fall out
81 Ok(handles
.into
_iter
().map(|h
| h
.join().unwrap
()).collect())
84 /// Calculate the md5 digest of some data on the given channel
85 fn md5_digest(rx
: Receiver
<Arc
<Vec
<u8>>>) -> JoinHandle
<(Algorithm
, Vec
<u8>)> {
86 thread
::spawn(move || {
87 let mut md5
= Md5
::new();
88 while let Ok(chunk
) = rx
.recv() {
91 let mut result
= [0; 16];
92 md5
.result(&mut result
);
93 (Algorithm
::Md5
, result
.to_vec())
97 /// Calculate the sha1 digest of some data on the given channel
98 fn sha1_digest(rx
: Receiver
<Arc
<Vec
<u8>>>) -> JoinHandle
<(Algorithm
, Vec
<u8>)> {
99 thread
::spawn(move || {
100 let mut sha1
= Sha1
::new();
101 while let Ok(chunk
) = rx
.recv() {
104 let mut result
= [0; 20];
105 sha1
.result(&mut result
);
106 (Algorithm
::Sha1
, result
.to_vec())
110 /// Calculate the sha256 digest of some data on the given channel
111 fn sha256_digest(rx
: Receiver
<Arc
<Vec
<u8>>>) -> JoinHandle
<(Algorithm
, Vec
<u8>)> {
112 thread
::spawn(move || {
113 let mut sha256
= Sha256
::new();
114 while let Ok(chunk
) = rx
.recv() {
115 sha256
.input(&chunk
);
117 let mut result
= [0; 32];
118 sha256
.result(&mut result
);
119 (Algorithm
::Sha256
, result
.to_vec())
128 static SMALL_DATA
: [u8; 10] = [b'A'
; 10];
129 // python3 -c 'print ("A"*10, end="", flush=True)' | md5sum
130 static SMALL_DATA_MD5
: &str = "16c52c6e8326c071da771e66dc6e9e57";
131 // python3 -c 'print ("A"*10, end="", flush=True)' | sha1sum
132 static SMALL_DATA_SHA1
: &str = "c71613a7386fd67995708464bf0223c0d78225c4";
133 // python3 -c 'print ("A"*10, end="", flush=True)' | sha256sum
134 static SMALL_DATA_SHA256
: &str =
135 "1d65bf29403e4fb1767522a107c827b8884d16640cf0e3b18c4c1dd107e0d49d";
137 static LARGE_DATA
: [u8; 1_000_000] = [b'B'
; 1_000_000];
138 // python3 -c 'print ("B"*1000000, end="", flush=True)' | md5sum
139 static LARGE_DATA_MD5
: &str = "9171f6d67a87ca649a702434a03458a1";
140 // python3 -c 'print ("B"*1000000, end="", flush=True)' | sha1sum
141 static LARGE_DATA_SHA1
: &str = "cfae4cebfd01884111bdede7cf983626bb249c94";
142 // python3 -c 'print ("B"*1000000, end="", flush=True)' | sha256sum
143 static LARGE_DATA_SHA256
: &str =
144 "b9193853f7798e92e2f6b82eda336fa7d6fc0fa90fdefe665f372b0bad8cdf8c";
146 fn verify_digest(alg
: Algorithm
, data
: &'
static [u8], hash
: &str) {
147 let reader
= Cursor
::new(&*data
);
148 let digests
= create_digests(&[alg
], Box
::new(reader
)).unwrap
();
149 assert_eq
!(digests
.len(), 1);
150 assert_eq
!(digests
[0], (alg
, hex
::decode(hash
).unwrap
()));
153 /// Assert that digests for all algorithms are calculated correctly for a small piece
154 /// of test data (single block).
157 verify_digest(Algorithm
::Md5
, &SMALL_DATA
, SMALL_DATA_MD5
);
158 verify_digest(Algorithm
::Sha1
, &SMALL_DATA
, SMALL_DATA_SHA1
);
159 verify_digest(Algorithm
::Sha256
, &SMALL_DATA
, SMALL_DATA_SHA256
);
162 /// Assert that digests for all algorithms are calculated correctly for a large piece
163 /// of test data. For our purposes, "large" means that it spans several of the 64 KB
164 /// blocks used to break up the input processing. Using one million bytes instead of
165 /// 1 MiB means that the final block will be slightly smaller than the others.
168 verify_digest(Algorithm
::Md5
, &LARGE_DATA
, LARGE_DATA_MD5
);
169 verify_digest(Algorithm
::Sha1
, &LARGE_DATA
, LARGE_DATA_SHA1
);
170 verify_digest(Algorithm
::Sha256
, &LARGE_DATA
, LARGE_DATA_SHA256
);