Handle pre-epoch file timestamps consistently and record Rust audit

This commit is contained in:
karl committed 2026-10-13 17:02:16 +02:00
1 parent 59b2fc39b2
commit 14adbbc616
7 files changed
+262 -46

No files matched your search

+80 -28
View File
@@ -778,34 +778,30 @@ impl Store {
))
})?;
if let (Some(size), Some(mtime)) = (size, mtime) {
let valid = *filesystem_validity
.entry(result.path.clone())
.or_insert_with(|| {
#[cfg(test)]
MEDIA_STAT_CALLS.with(|count| count.set(count.get() + 1));
let validation_span = trace.begin(
"semantic.parent_filesystem_validation",
materialization_span,
);
let valid =
std::fs::metadata(&result.path)
.ok()
.is_some_and(|metadata| {
metadata.is_file()
&& i64::try_from(metadata.len()).ok() == Some(size)
&& metadata
.modified()
.ok()
.and_then(|t| {
t.duration_since(std::time::UNIX_EPOCH).ok()
})
.and_then(|d| i64::try_from(d.as_nanos()).ok())
== Some(mtime)
});
trace.counter(validation_span, "valid", u64::from(valid));
trace.end(validation_span, Default::default());
valid
});
let valid =
*filesystem_validity
.entry(result.path.clone())
.or_insert_with(|| {
#[cfg(test)]
MEDIA_STAT_CALLS.with(|count| count.set(count.get() + 1));
let validation_span = trace.begin(
"semantic.parent_filesystem_validation",
materialization_span,
);
let valid =
std::fs::metadata(&result.path)
.ok()
.is_some_and(|metadata| {
metadata.is_file()
&& i64::try_from(metadata.len()).ok() == Some(size)
&& metadata.modified().ok().and_then(|t| {
searchi_media::unix_time_ns(t).ok()
}) == Some(mtime)
});
trace.counter(validation_span, "valid", u64::from(valid));
trace.end(validation_span, Default::default());
valid
});
if !valid {
continue;
}
@@ -1141,6 +1137,62 @@ mod tests {
Ok(())
}
#[test]
fn pre_epoch_media_parent_remains_eligible_for_semantic_results() -> Result<()> {
let directory = tempfile::tempdir()?;
let root = directory.path().join("media");
std::fs::create_dir(&root)?;
let path = root.join("archived.mp4");
std::fs::write(&path, b"bounded media identity fixture")?;
std::fs::File::options()
.write(true)
.open(&path)?
.set_times(
std::fs::FileTimes::new()
.set_modified(std::time::UNIX_EPOCH - std::time::Duration::from_secs(1)),
)?;
let db = directory.path().join("index.db");
let mut lexical = SearchIndex::open(&db)?;
lexical.index_folder(&root)?;
let mut store = Store::open(&db, DIMENSIONS)?;
let generation = store.begin("pre-epoch-media-proof")?;
let mut query = vec![0.; DIMENSIONS];
query[0] = 1.;
let (id, indexed_path, _, _) = store.media_batch(0)?.remove(0);
let identity = searchi_media::MediaIdentity::read(Path::new(&indexed_path), 1024)?;
assert_eq!(identity.modified_ns, -1_000_000_000);
store.insert_media(generation, id, 0, &identity, "video", 0, &query)?;
store.activate(generation)?;
let mut trace = TraceCollector::new(MetricsMode::Basic);
let results = store.search_origin_traced(
&query,
"pre-epoch-media-proof",
10,
&SearchOptions::default(),
None,
&mut trace,
None,
)?;
assert_eq!(results.len(), 1);
assert_eq!(results[0].path, indexed_path);
// Timestamp changes still invalidate the prepared media source.
std::fs::File::options()
.write(true)
.open(&path)?
.set_times(std::fs::FileTimes::new().set_modified(std::time::UNIX_EPOCH))?;
let results = store.search_origin_traced(
&query,
"pre-epoch-media-proof",
10,
&SearchOptions::default(),
None,
&mut trace,
None,
)?;
assert!(results.is_empty());
Ok(())
}
#[test]
fn many_video_windows_do_not_crowd_out_other_parent_files() -> Result<()> {
// Geometry verifies grouped retrieval, not model inference or codecs.