1
Fork 0
mirror of https://github.com/RGBCube/uutils-coreutils synced 2025-07-29 12:07:46 +00:00

checksum: remove ALGO_BASED_REGEX (non base64) as its not useful anymore and introduce LineFormat struct

This commit is contained in:
Dorian Péron 2024-12-07 02:38:00 +01:00 committed by Dorian Péron
parent f4e5dc2e0f
commit 567bbc5f3c

View file

@ -421,8 +421,7 @@ pub fn detect_algo(algo: &str, length: Option<usize>) -> UResult<HashAlgorithm>
// algo must be uppercase or b (for blake2b) // algo must be uppercase or b (for blake2b)
// 2. <checksum> [* ]<filename> // 2. <checksum> [* ]<filename>
// 3. <checksum> [*]<filename> (only one space) // 3. <checksum> [*]<filename> (only one space)
const ALGO_BASED_REGEX: &str = r"^\s*\\?(?P<algo>(?:[A-Z0-9]+|BLAKE2b))(?:-(?P<bits>\d+))?\s?\((?P<filename>(?-u:.*))\)\s*=\s*(?P<checksum>[a-fA-F0-9]+)$"; const ALGO_BASED_REGEX: &str = r"^\s*\\?(?P<algo>(?:[A-Z0-9]+|BLAKE2b))(?:-(?P<bits>\d+))?\s?\((?P<filename>(?-u:.*))\)\s*=\s*(?P<checksum>[A-Za-z0-9+/]+={0,2})$";
const ALGO_BASED_REGEX_BASE64: &str = r"^\s*\\?(?P<algo>(?:[A-Z0-9]+|BLAKE2b))(?:-(?P<bits>\d+))?\s?\((?P<filename>(?-u:.*))\)\s*=\s*(?P<checksum>[A-Za-z0-9+/]+={0,2})$";
const DOUBLE_SPACE_REGEX: &str = r"^(?P<checksum>[a-fA-F0-9]+)\s{2}(?P<filename>(?-u:.*))$"; const DOUBLE_SPACE_REGEX: &str = r"^(?P<checksum>[a-fA-F0-9]+)\s{2}(?P<filename>(?-u:.*))$";
@ -433,7 +432,23 @@ lazy_static! {
static ref R_ALGO_BASED: Regex = Regex::new(ALGO_BASED_REGEX).unwrap(); static ref R_ALGO_BASED: Regex = Regex::new(ALGO_BASED_REGEX).unwrap();
static ref R_DOUBLE_SPACE: Regex = Regex::new(DOUBLE_SPACE_REGEX).unwrap(); static ref R_DOUBLE_SPACE: Regex = Regex::new(DOUBLE_SPACE_REGEX).unwrap();
static ref R_SINGLE_SPACE: Regex = Regex::new(SINGLE_SPACE_REGEX).unwrap(); static ref R_SINGLE_SPACE: Regex = Regex::new(SINGLE_SPACE_REGEX).unwrap();
static ref R_ALGO_BASED_BASE_64: Regex = Regex::new(ALGO_BASED_REGEX_BASE64).unwrap(); }
#[derive(Debug, PartialEq, Eq, Clone, Copy)]
enum LineFormat {
AlgoBased,
SingleSpace,
DoubleSpace,
}
impl LineFormat {
fn to_regex(self) -> &'static Regex {
match self {
LineFormat::AlgoBased => &R_ALGO_BASED,
LineFormat::SingleSpace => &R_SINGLE_SPACE,
LineFormat::DoubleSpace => &R_DOUBLE_SPACE,
}
}
} }
/// Hold the data extracted from a checksum line. /// Hold the data extracted from a checksum line.
@ -443,34 +458,41 @@ struct LineInfo {
checksum: String, checksum: String,
filename: Vec<u8>, filename: Vec<u8>,
regex: &'static Regex, format: LineFormat,
} }
impl LineInfo { impl LineInfo {
fn parse(s: impl AsRef<OsStr>, cached_regex: &mut Option<&'static Regex>) -> Option<Self> { /// Returns a `LineInfo` parsed from a checksum line.
let regexes: &[(&'static Regex, bool)] = &[ /// The function will run 3 regexes against the line and select the first one that matches
(&R_ALGO_BASED, true), /// to populate the fields of the struct.
(&R_DOUBLE_SPACE, false), /// However, there is a catch to handle regarding the handling of `cached_regex`.
(&R_SINGLE_SPACE, false), /// In case of non-algo-based regex, if `cached_regex` is Some, it must take the priority
(&R_ALGO_BASED_BASE_64, true), /// over the detected regex. Otherwise, we must set it the the detected regex.
/// This specific behavior is emphasized by the test
/// `test_hashsum::test_check_md5sum_only_one_space`.
fn parse(s: impl AsRef<OsStr>, cached_regex: &mut Option<LineFormat>) -> Option<Self> {
let regexes: &[(&'static Regex, LineFormat)] = &[
(&R_ALGO_BASED, LineFormat::AlgoBased),
(&R_DOUBLE_SPACE, LineFormat::DoubleSpace),
(&R_SINGLE_SPACE, LineFormat::SingleSpace),
]; ];
let line_bytes = os_str_as_bytes(s.as_ref()).expect("UTF-8 decoding failed"); let line_bytes = os_str_as_bytes(s.as_ref()).expect("UTF-8 decoding failed");
for (regex, algo_based) in regexes { for (regex, format) in regexes {
if !regex.is_match(line_bytes) { if !regex.is_match(line_bytes) {
continue; continue;
} }
let mut r = *regex; let mut r = *regex;
if !algo_based { if *format != LineFormat::AlgoBased {
// The cached regex ensures that when processing non-algo based regexes, // The cached regex ensures that when processing non-algo based regexes,
// its cannot be changed (can't have single and double space regexes // it cannot be changed (can't have single and double space regexes
// used in the same file). // used in the same file).
if cached_regex.is_some() { if cached_regex.is_some() {
r = cached_regex.unwrap(); r = cached_regex.unwrap().to_regex();
} else { } else {
*cached_regex = Some(r); *cached_regex = Some(*format);
} }
} }
@ -485,23 +507,13 @@ impl LineInfo {
.map(|m| match_to_string(m).parse::<usize>().unwrap()), .map(|m| match_to_string(m).parse::<usize>().unwrap()),
checksum: caps.name("checksum").map(match_to_string).unwrap(), checksum: caps.name("checksum").map(match_to_string).unwrap(),
filename: caps.name("filename").map(|m| m.as_bytes().into()).unwrap(), filename: caps.name("filename").map(|m| m.as_bytes().into()).unwrap(),
regex: r, format: *format,
}); });
} }
} }
None None
} }
#[inline]
fn is_algo_based(&self) -> bool {
self.algo_name.is_some()
}
#[inline]
fn regex_str(&self) -> &str {
self.regex.as_str()
}
} }
fn get_filename_for_output(filename: &OsStr, input_is_stdin: bool) -> String { fn get_filename_for_output(filename: &OsStr, input_is_stdin: bool) -> String {
@ -730,14 +742,16 @@ fn process_algo_based_line(
/// Check a digest checksum with non-algo based pre-treatment. /// Check a digest checksum with non-algo based pre-treatment.
fn process_non_algo_based_line( fn process_non_algo_based_line(
i: usize, line_number: usize,
line_info: &LineInfo, line_info: &LineInfo,
cli_algo_name: &str, cli_algo_name: &str,
cli_algo_length: Option<usize>, cli_algo_length: Option<usize>,
opts: ChecksumOptions, opts: ChecksumOptions,
) -> Result<(), LineCheckError> { ) -> Result<(), LineCheckError> {
let mut filename_to_check = line_info.filename.as_slice(); let mut filename_to_check = line_info.filename.as_slice();
if filename_to_check.starts_with(b"*") && i == 0 && line_info.regex_str() == SINGLE_SPACE_REGEX if filename_to_check.starts_with(b"*")
&& line_number == 0
&& line_info.format == LineFormat::SingleSpace
{ {
// Remove the leading asterisk if present - only for the first line // Remove the leading asterisk if present - only for the first line
filename_to_check = &filename_to_check[1..]; filename_to_check = &filename_to_check[1..];
@ -774,7 +788,7 @@ fn process_checksum_line(
cli_algo_name: Option<&str>, cli_algo_name: Option<&str>,
cli_algo_length: Option<usize>, cli_algo_length: Option<usize>,
opts: ChecksumOptions, opts: ChecksumOptions,
cached_regex: &mut Option<&'static Regex>, cached_regex: &mut Option<LineFormat>,
) -> Result<(), LineCheckError> { ) -> Result<(), LineCheckError> {
let line_bytes = os_str_as_bytes(line)?; let line_bytes = os_str_as_bytes(line)?;
@ -786,7 +800,7 @@ fn process_checksum_line(
// Use `LineInfo` to extract the data of a line. // Use `LineInfo` to extract the data of a line.
// Then, depending on its format, apply a different pre-treatment. // Then, depending on its format, apply a different pre-treatment.
if let Some(line_info) = LineInfo::parse(line, cached_regex) { if let Some(line_info) = LineInfo::parse(line, cached_regex) {
if line_info.is_algo_based() { if line_info.format == LineFormat::AlgoBased {
process_algo_based_line(&line_info, cli_algo_name, opts) process_algo_based_line(&line_info, cli_algo_name, opts)
} else if let Some(cli_algo) = cli_algo_name { } else if let Some(cli_algo) = cli_algo_name {
// If we match a non-algo based regex, we expect a cli argument // If we match a non-algo based regex, we expect a cli argument
@ -1284,23 +1298,21 @@ mod tests {
let line_algo_based = let line_algo_based =
OsString::from("MD5 (example.txt) = d41d8cd98f00b204e9800998ecf8427e"); OsString::from("MD5 (example.txt) = d41d8cd98f00b204e9800998ecf8427e");
let line_info = LineInfo::parse(&line_algo_based, &mut cached_regex).unwrap(); let line_info = LineInfo::parse(&line_algo_based, &mut cached_regex).unwrap();
assert!(line_info.is_algo_based());
assert_eq!(line_info.algo_name.as_deref(), Some("MD5")); assert_eq!(line_info.algo_name.as_deref(), Some("MD5"));
assert!(line_info.algo_bit_len.is_none()); assert!(line_info.algo_bit_len.is_none());
assert_eq!(line_info.filename, b"example.txt"); assert_eq!(line_info.filename, b"example.txt");
assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e"); assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e");
assert_eq!(line_info.regex_str(), ALGO_BASED_REGEX); assert_eq!(line_info.format, LineFormat::AlgoBased);
assert!(cached_regex.is_none()); assert!(cached_regex.is_none());
// Test double-space regex // Test double-space regex
let line_double_space = OsString::from("d41d8cd98f00b204e9800998ecf8427e example.txt"); let line_double_space = OsString::from("d41d8cd98f00b204e9800998ecf8427e example.txt");
let line_info = LineInfo::parse(&line_double_space, &mut cached_regex).unwrap(); let line_info = LineInfo::parse(&line_double_space, &mut cached_regex).unwrap();
assert!(!line_info.is_algo_based());
assert!(line_info.algo_name.is_none()); assert!(line_info.algo_name.is_none());
assert!(line_info.algo_bit_len.is_none()); assert!(line_info.algo_bit_len.is_none());
assert_eq!(line_info.filename, b"example.txt"); assert_eq!(line_info.filename, b"example.txt");
assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e"); assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e");
assert_eq!(line_info.regex_str(), DOUBLE_SPACE_REGEX); assert_eq!(line_info.format, LineFormat::DoubleSpace);
assert!(cached_regex.is_some()); assert!(cached_regex.is_some());
cached_regex = None; cached_regex = None;
@ -1308,12 +1320,11 @@ mod tests {
// Test single-space regex // Test single-space regex
let line_single_space = OsString::from("d41d8cd98f00b204e9800998ecf8427e example.txt"); let line_single_space = OsString::from("d41d8cd98f00b204e9800998ecf8427e example.txt");
let line_info = LineInfo::parse(&line_single_space, &mut cached_regex).unwrap(); let line_info = LineInfo::parse(&line_single_space, &mut cached_regex).unwrap();
assert!(!line_info.is_algo_based());
assert!(line_info.algo_name.is_none()); assert!(line_info.algo_name.is_none());
assert!(line_info.algo_bit_len.is_none()); assert!(line_info.algo_bit_len.is_none());
assert_eq!(line_info.filename, b"example.txt"); assert_eq!(line_info.filename, b"example.txt");
assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e"); assert_eq!(line_info.checksum, "d41d8cd98f00b204e9800998ecf8427e");
assert_eq!(line_info.regex_str(), SINGLE_SPACE_REGEX); assert_eq!(line_info.format, LineFormat::SingleSpace);
assert!(cached_regex.is_some()); assert!(cached_regex.is_some());
cached_regex = None; cached_regex = None;
@ -1328,7 +1339,7 @@ mod tests {
OsString::from(" MD5 (example.txt) = d41d8cd98f00b204e9800998ecf8427e"); OsString::from(" MD5 (example.txt) = d41d8cd98f00b204e9800998ecf8427e");
let res = LineInfo::parse(&line_algo_based_leading_space, &mut cached_regex); let res = LineInfo::parse(&line_algo_based_leading_space, &mut cached_regex);
assert!(res.is_some()); assert!(res.is_some());
assert_eq!(res.unwrap().regex_str(), ALGO_BASED_REGEX); assert_eq!(line_info.format, LineFormat::AlgoBased);
assert!(cached_regex.is_none()); assert!(cached_regex.is_none());
// Test trailing space after checksum line (should fail) // Test trailing space after checksum line (should fail)