eblume/kingfisher
1
0
Fork 0
forked from mirrors/kingfisher
Code Pull requests Activity Actions

fixing issues reported by code review

Browse source
This commit is contained in:
Mick Grove 2025-08-17 18:02:36 -07:00
commit d3b85ed370
2 changed files with 22 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

29
src/content_type.rs
View file

@ -1,7 +1,22 @@
// Requires: tokei = "12" in Cargo.toml
use once_cell::sync::Lazy;
use std::path::Path;
use tokei::LanguageType;
// Precompute all (shebang_prefix_bytes, language) pairs once.
// Sort longest-first so more specific shebangs win.
static SHEBANG_PREFIXES: Lazy<Vec<(&'static [u8], LanguageType)>> = Lazy::new(|| {
let mut v = Vec::new();
for &lang in LanguageType::list() {
for &sb in lang.shebangs() {
v.push((sb.as_bytes(), lang));
}
}
// Longest prefix first to prefer e.g. "#!/usr/bin/env python3" over "#!/usr/bin/env python"
v.sort_by(|a, b| b.0.len().cmp(&a.0.len()));
v
});
/// The type of content detected.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum ContentType {
@ -127,16 +142,12 @@ impl ContentInspector {
}
}
// 3) Shebang detection (in-memory): compare the first line to known shebangs.
// 3) Shebang detection (in-memory): match by longest prefix, byte-wise (no UTF-8 needed).
if let Some(first_line) = content.split(|&b| b == b'\n').next() {
if first_line.starts_with(b"#!") {
if let Ok(line) = std::str::from_utf8(first_line) {
for &lang in LanguageType::list() {
for &sb in lang.shebangs() {
if line.starts_with(sb) {
return Some(lang.name().to_string());
}
}
for (prefix, lang) in SHEBANG_PREFIXES.iter() {
if first_line.starts_with(prefix) {
return Some(lang.name().to_string());
}
}
}

4
src/scanner/enumerate.rs
View file

@ -455,8 +455,8 @@ impl<'a> rayon::iter::ParallelIterator for GitRepoResultIter<'a> {
let origin = OriginSet::try_from_iter(md.first_seen.iter().map(|e| {
Origin::from_git_repo_with_first_commit(
Arc::clone(&repo_path),
Arc::clone(&e.commit_metadata), // ← clone Arc
String::from_utf8_lossy(&e.path).into_owned(),
Arc::clone(&e.commit_metadata),
String::from_utf8_lossy(&e.path).to_string(),
)
}))
.unwrap_or_else(|| Origin::from_git_repo(Arc::clone(&repo_path)).into());