fix(interchange): avoid UTF-8 slice panics in file URLs

Use a checked localhost prefix probe so Unicode file URLs retain their media paths without slicing inside a UTF-8 character. Add public converter, resolver and EDL import regressions.

Fixes #1591

Co-authored-by: Codex <noreply@openai.com>

crates/interchange/README.md

@@ -24,6 +24,10 @@ read off files it exports, never from the application itself.
 
 The other formats' editions are listed in the module docs (`edl`, `fcp7`, `fcpxml`, `otio`, `ale`).
 
+**Invariant — Unicode media URLs:** file URL prefix probes must not panic when their byte boundary
+falls inside a UTF-8 character. Timeline importers use the shared converter; `tests/file_urls.rs`
+checks expected Unicode paths through the public converter, resolver and EDL importer.
+
 ## AAF and OMF
 
 `comp` holds a format-neutral mob-style model (compositions of slots whose sequences of fillers,

crates/interchange/src/common.rs

@@ -65,7 +65,7 @@ pub fn file_url_to_path(url: &str) -> String {
         return url.to_string();
     }
     let rest = &url[7..];
-    let rest = if rest.len() >= 9 && rest[..9].eq_ignore_ascii_case("localhost") { &rest[9..] } else { rest };
+    let rest = if rest.get(..9).is_some_and(|s| s.eq_ignore_ascii_case("localhost")) { &rest[9..] } else { rest };
     let decoded = percent_decode(rest);
     // file:///C:/x -> C:/x
     if decoded.len() >= 3 && decoded.starts_with('/') && is_windows_abs(&decoded[1..]) {

crates/interchange/tests/file_urls.rs

@@ -0,0 +1,45 @@
+//! Media URL handling at the public interchange boundary.
+
+use filmcraft_interchange::{Format, file_url_to_path, import, resolve_path};
+use filmcraft_project::MediaRef;
+
+#[test]
+fn unicode_file_urls_do_not_panic_or_lose_the_path() {
+    for (url, path) in [("file://srv/日本語.mov", "//srv/日本語.mov"), ("file://localhost/media/漢字.mov", "/media/漢字.mov")] {
+        let parsed = std::panic::catch_unwind(|| file_url_to_path(url));
+        assert!(parsed.is_ok(), "file URL panicked: {url}");
+        assert_eq!(parsed.unwrap(), path);
+        assert_eq!(resolve_path(url, Some("/project")), path);
+    }
+    for path in ["/media/漢字.mov", "/日本語.mov", "//srv/日本語.mov"] {
+        let url = format!("file://{path}");
+        let parsed = std::panic::catch_unwind(|| file_url_to_path(&url));
+        assert!(parsed.is_ok(), "file URL panicked: {url}");
+        assert_eq!(parsed.unwrap(), path);
+        assert_eq!(resolve_path(&url, Some("/project")), path);
+    }
+}
+
+#[test]
+fn edl_import_keeps_unicode_file_url_media() {
+    let doc = "TITLE: Unicode media
+FCM: NON-DROP FRAME
+001 AX V C 00:00:00:00 00:00:01:00 00:00:00:00 00:00:01:00
+* SOURCE FILE: file:///日本語.mov
+";
+    let parsed = std::panic::catch_unwind(|| import(doc.as_bytes(), Format::Edl, None));
+    assert!(parsed.is_ok(), "importing a Unicode file URL panicked");
+    let (imported, report) = parsed.unwrap().unwrap();
+    let paths: Vec<_> = imported
+        .project
+        .items
+        .values()
+        .filter_map(|item| item.as_media())
+        .filter_map(|media| match &media.media {
+            MediaRef::File { path } => Some(path.as_str()),
+            _ => None,
+        })
+        .collect();
+    assert_eq!(paths, ["/日本語.mov"]);
+    assert!(!report.has_warnings(), "{report}");
+}