};
#[cfg(feature = "paste")]
use clipboard::{ClipboardContext, ClipboardProvider};
-use regex::Regex;
use std::fs::File;
use std::io;
use std::io::prelude::*;
))
}
+fn try_parse_hash(s: &str) -> Option<(Algorithm, Vec<u8>)> {
+ let bytes = match hex::decode(s.trim()) {
+ Ok(bytes) => bytes,
+ _ => return None,
+ };
+ let alg = match Algorithm::from_len(bytes.len()) {
+ Ok(alg) => alg,
+ _ => return None,
+ };
+ Some((alg, bytes))
+}
+
fn read_raw_candidate_from_file(line: &str, path: &PathBuf) -> Option<CandidateHashes> {
- // It is a little sad to use a dynamic regex in an otherwise nice Rust program
- // These deserve to be replaced with a good old fashioned static parser
- // But let's be honest: the impact is negligible
- let re = Regex::new(r"^([[:xdigit:]]{32}|[[:xdigit:]]{40}|[[:xdigit:]]{64})$").unwrap();
- if re.is_match(line) {
- // These should both always succeed due to the matching
- let bytes = match hex::decode(line) {
- Ok(bytes) => bytes,
- _ => return None,
- };
- let alg = match Algorithm::from_len(bytes.len()) {
- Ok(alg) => alg,
- _ => return None,
- };
- return Some(CandidateHashes {
- alg,
- source: VerificationSource::RawFile(path.clone()),
- hashes: vec![CandidateHash {
- bytes,
- filename: None,
- }],
- });
- }
- None
+ let (alg, bytes) = try_parse_hash(line)?;
+ Some(CandidateHashes {
+ alg,
+ source: VerificationSource::RawFile(path.clone()),
+ hashes: vec![CandidateHash {
+ bytes,
+ filename: None,
+ }],
+ })
}
-fn read_coreutils_digests_from_file<I>(lines: I, path: &PathBuf) -> Option<CandidateHashes>
+fn read_coreutils_digests_from_file<I, S>(lines: I, path: &PathBuf) -> Option<CandidateHashes>
where
- I: Iterator<Item = io::Result<String>>,
+ I: Iterator<Item = io::Result<S>>,
+ S: AsRef<str>,
{
- let re = Regex::new(
- r"^(?P<hash>([[:xdigit:]]{32}|[[:xdigit:]]{40}|[[:xdigit:]]{64})) .(?P<filename>.+)$",
- )
- .unwrap();
-
let mut hashes = vec![];
let mut alg: Option<Algorithm> = None;
for l in lines {
if let Ok(l) = l {
- let l = l.trim();
+ let l = l.as_ref().trim();
// Allow (ignore) blank lines
if l.is_empty() {
continue;
}
- // If we can capture a valid line, use it
- if let Some(captures) = re.captures(&l) {
- let hash = &captures["hash"];
- let filename = &captures["filename"];
- // Decode the hex and algorithm for this line
- let line_bytes = match hex::decode(hash) {
- Ok(bytes) => bytes,
- _ => return None,
- };
- let line_alg = match Algorithm::from_len(line_bytes.len()) {
- Ok(alg) => alg,
- _ => return None,
- };
- if alg.is_some() && alg != Some(line_alg) {
- // Different algorithms in the same digest file are not supported
+ // Expected format
+ // <valid-hash><space><space-or-*><filename>
+ let (line_alg, bytes, filename) = match l
+ .find(' ')
+ .and_then(|space_pos| (l.get(0..space_pos)).zip(l.get(space_pos + 2..)))
+ .and_then(|(maybe_hash, filename)| {
+ try_parse_hash(maybe_hash).map(|(alg, bytes)| (alg, bytes, filename))
+ }) {
+ Some(t) => t,
+ None => {
+ // if we have a line with content we cannot parse, this is an error
return None;
- } else {
- // If we are the first line, we define the overall algorithm
- alg = Some(line_alg);
}
- // So far so good - create an entry for this line
- hashes.push(CandidateHash {
- bytes: line_bytes,
- filename: Some(filename.to_owned()),
- });
- } else {
- // But if we have a line with content we cannot parse, this is an error
+ };
+ if alg.is_some() && alg != Some(line_alg) {
+ // Different algorithms in the same digest file are not supported
return None;
+ } else {
+ // If we are the first line, we define the overall algorithm
+ alg = Some(line_alg);
}
+ // So far so good - create an entry for this line
+ hashes.push(CandidateHash {
+ bytes,
+ filename: Some(filename.to_owned()),
+ });
}
}
#[cfg(test)]
mod tests {
+ use std::path::Path;
+
use super::*;
#[test]
assert!(read_raw_candidate_from_file(*i, &example_path).is_none());
}
}
+
+ #[test]
+ fn test_read_shasums() {
+ let shasums = "4b91f7a387a6edd4a7c0afb2897f1ca968c9695b *cp
+ 75eb7420a9f5a260b04a3e8ad51e50f2838a17fc lel.txt
+
+ fe6c26d485a3573a1cb0ad0682f5105325a1905f shasums";
+ let lines = shasums.lines().map(|l| std::io::Result::Ok(l));
+ let path = Path::new("SHASUMS").to_owned();
+ let candidates = read_coreutils_digests_from_file(lines, &path);
+
+ assert_eq!(
+ candidates,
+ Some(CandidateHashes {
+ alg: Algorithm::Sha1,
+ hashes: vec![
+ CandidateHash {
+ bytes: hex::decode("4b91f7a387a6edd4a7c0afb2897f1ca968c9695b").unwrap(),
+ filename: Some("cp".to_owned()),
+ },
+ CandidateHash {
+ bytes: hex::decode("75eb7420a9f5a260b04a3e8ad51e50f2838a17fc").unwrap(),
+ filename: Some("lel.txt".to_owned()),
+ },
+ CandidateHash {
+ bytes: hex::decode("fe6c26d485a3573a1cb0ad0682f5105325a1905f").unwrap(),
+ filename: Some("shasums".to_owned()),
+ }
+ ],
+ source: VerificationSource::DigestsFile(path),
+ })
+ );
+ }
}