Marketplace: repo tree view and catalog parsing

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-09-27 08:48:31 -07:00
co-authored by Claude Opus 5.5
parent d419d0a6b4
commit 2e62728b06
4 changed files with 1181 additions and 0 deletions
+229
View File
@@ -0,0 +1,229 @@
//! A read-only view of a repository tree at one commit.
//!
//! The catalog parser only ever talks to [`TreeView`], so it is tested
//! against [`MemTree`] with no git involved, and runs in production against
//! [`GitTree`], which reads git objects straight out of the bare cache.
use std::collections::BTreeMap;
use sha2::{Digest, Sha256};
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum EntryKind {
File,
Dir,
Symlink,
/// Anything else git can hold (submodule commits). Never installable.
Other,
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct DirEntry {
pub name: String,
pub kind: EntryKind,
pub executable: bool,
}
pub trait TreeView {
/// Entries of the directory at `path` (`""` = root). `Ok(None)` if absent or not a dir.
fn list_dir(&self, path: &str) -> Result<Option<Vec<DirEntry>>, String>;
/// Contents of the regular file at `path`. `Ok(None)` if absent or not a file.
fn read_file(&self, path: &str) -> Result<Option<Vec<u8>>, String>;
/// Stable content id of the entry at `path`; `None` if absent.
fn entry_id(&self, path: &str) -> Result<Option<String>, String>;
}
/// Hex-encode `bytes`. Shared by [`MemTree`]'s content id (test-only) and
/// `catalog::item_fingerprint`'s plugin-entry hash (production), so there is
/// one hex formatter rather than two copies of the same `format!("{:02x}")`.
pub(crate) fn hex(bytes: &[u8]) -> String {
bytes.iter().map(|b| format!("{:02x}", b)).collect()
}
#[cfg(test)]
#[derive(Debug, Clone)]
enum MemNode {
File { data: Vec<u8>, executable: bool },
Symlink { target: String },
}
/// In-memory tree for tests: path → node. Directories are implied by paths.
#[cfg(test)]
#[derive(Debug, Clone, Default)]
pub struct MemTree {
nodes: BTreeMap<String, MemNode>,
}
#[cfg(test)]
impl MemTree {
pub fn new() -> Self {
Self::default()
}
pub fn file(mut self, path: &str, contents: &str) -> Self {
self.nodes.insert(
path.to_string(),
MemNode::File {
data: contents.as_bytes().to_vec(),
executable: false,
},
);
self
}
pub fn exec_file(mut self, path: &str, contents: &str) -> Self {
self.nodes.insert(
path.to_string(),
MemNode::File {
data: contents.as_bytes().to_vec(),
executable: true,
},
);
self
}
pub fn symlink(mut self, path: &str, target: &str) -> Self {
self.nodes.insert(
path.to_string(),
MemNode::Symlink {
target: target.to_string(),
},
);
self
}
fn is_dir(&self, path: &str) -> bool {
if path.is_empty() {
return true;
}
let prefix = format!("{}/", path);
self.nodes.keys().any(|k| k.starts_with(&prefix))
}
}
#[cfg(test)]
impl TreeView for MemTree {
fn list_dir(&self, path: &str) -> Result<Option<Vec<DirEntry>>, String> {
if self.nodes.contains_key(path) || !self.is_dir(path) {
return Ok(None);
}
let prefix = if path.is_empty() {
String::new()
} else {
format!("{}/", path)
};
let mut out: BTreeMap<String, DirEntry> = BTreeMap::new();
for (key, node) in &self.nodes {
let Some(rest) = key.strip_prefix(&prefix) else {
continue;
};
match rest.split_once('/') {
Some((dir, _)) => {
out.entry(dir.to_string()).or_insert(DirEntry {
name: dir.to_string(),
kind: EntryKind::Dir,
executable: false,
});
}
None => {
let (kind, executable) = match node {
MemNode::File { executable, .. } => (EntryKind::File, *executable),
MemNode::Symlink { .. } => (EntryKind::Symlink, false),
};
out.insert(
rest.to_string(),
DirEntry {
name: rest.to_string(),
kind,
executable,
},
);
}
}
}
Ok(Some(out.into_values().collect()))
}
fn read_file(&self, path: &str) -> Result<Option<Vec<u8>>, String> {
match self.nodes.get(path) {
Some(MemNode::File { data, .. }) => Ok(Some(data.clone())),
_ => Ok(None),
}
}
fn entry_id(&self, path: &str) -> Result<Option<String>, String> {
let mut hasher = Sha256::new();
let mut found = false;
let prefix = format!("{}/", path);
for (key, node) in &self.nodes {
if key != path && !key.starts_with(&prefix) {
continue;
}
found = true;
hasher.update(key.as_bytes());
hasher.update([0]);
match node {
MemNode::File { data, executable } => {
hasher.update([if *executable { b'x' } else { b'f' }]);
hasher.update(data);
}
MemNode::Symlink { target } => {
hasher.update(b"l");
hasher.update(target.as_bytes());
}
}
hasher.update([0]);
}
Ok(found.then(|| hex(&hasher.finalize())))
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn mem_tree_lists_files_dirs_and_symlinks() {
let t = MemTree::new()
.file("agents/a.md", "x")
.exec_file("hooks/h/run.sh", "#!/bin/sh")
.symlink("agents/link.md", "a.md");
let root = t.list_dir("").unwrap().unwrap();
assert_eq!(
root.iter()
.map(|e| (e.name.as_str(), e.kind))
.collect::<Vec<_>>(),
vec![("agents", EntryKind::Dir), ("hooks", EntryKind::Dir)]
);
let agents = t.list_dir("agents").unwrap().unwrap();
assert_eq!(agents[1].kind, EntryKind::Symlink);
let hook = t.list_dir("hooks/h").unwrap().unwrap();
assert!(hook[0].executable);
assert_eq!(t.list_dir("agents/a.md").unwrap(), None);
assert_eq!(t.list_dir("missing").unwrap(), None);
assert_eq!(t.read_file("agents/a.md").unwrap().unwrap(), b"x");
assert_eq!(t.read_file("agents").unwrap(), None);
}
#[test]
fn mem_tree_entry_id_changes_only_with_content() {
let a = MemTree::new()
.file("skills/s/SKILL.md", "one")
.file("agents/x.md", "x");
let b = MemTree::new()
.file("skills/s/SKILL.md", "one")
.file("agents/x.md", "changed");
let c = MemTree::new()
.file("skills/s/SKILL.md", "two")
.file("agents/x.md", "x");
assert_eq!(
a.entry_id("skills/s").unwrap(),
b.entry_id("skills/s").unwrap()
);
assert_ne!(
a.entry_id("skills/s").unwrap(),
c.entry_id("skills/s").unwrap()
);
assert_eq!(a.entry_id("nope").unwrap(), None);
}
}