actions/typos
1
0
Fork 0
mirror of https://github.com/crate-ci/typos.git synced 2024-11-22 17:11:07 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #330 from nmathewson/flexible_hash_literals

Browse source 
Ignore hexadecimal "hashes" of length 32 or greater.
This commit is contained in:
Ed Page 2021-08-20 12:11:06 -05:00 committed by GitHub
commit 44d9828330
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 53 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

59
crates/typos/src/tokens.rs
View file

@ -264,13 +264,20 @@ mod parser {
<T as nom::InputIter>::Item: AsChar + Copy, <T as nom::InputIter>::Item: AsChar + Copy,
{ {
// Size considerations: // Size considerations:
// - sha-1 is git's original hash // - 40 characters holds for a SHA-1 hash from older Git versions.
// - sha-256 is git's new hash // - 64 characters holds for a SHA-256 hash from newer Git versions.
// - Git hashes can be abbreviated but we need a good abbreviation that won't be mistaken // - Git allows abbreviated hashes, but we need a good abbreviation
// for a variable name // that won't be mistaken for a variable name.
const SHA_1_MAX: usize = 40; // - Through experimentation we've found that there is almost
const SHA_256_MAX: usize = 64; // never any actual text inside a hex string of 32 characters
take_while_m_n(SHA_1_MAX, SHA_256_MAX, is_lower_hex_digit)(input) // or more.
const IGNORE_HEX_MIN: usize = 32;
const IGNORE_HEX_MAX: usize = usize::MAX;
alt((
take_while_m_n(IGNORE_HEX_MIN, IGNORE_HEX_MAX, is_lower_hex_digit),
take_while_m_n(IGNORE_HEX_MIN, IGNORE_HEX_MAX, is_upper_hex_digit),
))(input)
} }
fn base64_literal<T>(input: T) -> IResult<T, T> fn base64_literal<T>(input: T) -> IResult<T, T>
@ -443,6 +450,12 @@ mod parser {
('a'..='f').contains(&c) || ('0'..='9').contains(&c) ('a'..='f').contains(&c) || ('0'..='9').contains(&c)
} }
#[inline]
fn is_upper_hex_digit(i: impl AsChar + Copy) -> bool {
let c = i.as_char();
('A'..='F').contains(&c) || ('0'..='9').contains(&c)
}
#[inline] #[inline]
fn is_base64_digit(i: impl AsChar + Copy) -> bool { fn is_base64_digit(i: impl AsChar + Copy) -> bool {
let c = i.as_char(); let c = i.as_char();
@ -905,16 +918,40 @@ mod test {
fn tokenize_ignore_hash() { fn tokenize_ignore_hash() {
let parser = TokenizerBuilder::new().build(); let parser = TokenizerBuilder::new().build();
let input = "Hello 485865fd0412e40d041e861506bb3ac11a3a91e3 World"; for (hashlike, is_ignored) in [
let expected: Vec<Identifier> = vec![ // A SHA-1 output, in lower case.
("485865fd0412e40d041e861506bb3ac11a3a91e3", true),
// A SHA-1 output, in mixed case: Not a ignored.
("485865fd0412e40d041E861506BB3AC11A3A91E3", false),
// A SHA-256 output, in upper case.
("E3B0C44298FC1C149AFBF4C8996FB92427AE41E4649B934CA495991B7852B855", true),
// A SHA-512 output, in lower case.
("cf83e1357eefb8bdf1542850d66d8007d620e4050b5715dc83f4a921d36ce9ce47d0d13c5d85f2b0ff8318d2877eec2f63b931bd47417a81a538327af927da3e", true),
// An MD5 (deprecated) output, in upper case.
("D41D8CD98F00B204E9800998ECF8427E", true),
// A 31-character hexadecimal string: too short to be a hash.
("D41D8CD98F00B204E9800998ECF8427", false),
// A 40-character string, but with non-hex characters (in
// several positions.)
("Z85865fd0412e40d041e861506bb3ac11a3a91e3", false),
("485865fd04Z2e40d041e861506bb3ac11a3a91e3", false),
("485865fd0412e40d041e8Z1506bb3ac11a3a91e3", false),
("485865fd0412e40d041e861506bb3ac11a3a91eZ", false),
] {
let input = format!("Hello {} World", hashlike);
let mut expected: Vec<Identifier> = vec![
Identifier::new_unchecked("Hello", Case::None, 0), Identifier::new_unchecked("Hello", Case::None, 0),
Identifier::new_unchecked("World", Case::None, 47), Identifier::new_unchecked("World", Case::None, 7+hashlike.len()),
]; ];
if ! is_ignored {
expected.insert(1, Identifier::new_unchecked(hashlike, Case::None, 6));
}
let actual: Vec<_> = parser.parse_bytes(input.as_bytes()).collect(); let actual: Vec<_> = parser.parse_bytes(input.as_bytes()).collect();
assert_eq!(expected, actual); assert_eq!(expected, actual);
let actual: Vec<_> = parser.parse_str(input).collect(); let actual: Vec<_> = parser.parse_str(&input).collect();
assert_eq!(expected, actual); assert_eq!(expected, actual);
} }
}
#[test] #[test]
fn tokenize_ignore_base64() { fn tokenize_ignore_base64() {