fix(fingerprint): explicit reason rather than "stale; unknown reason" (#15617)

### What does this PR try to resolve?

This was discovered during playing with rmeta reuse between cargo-check
and cargo-build.

An rmeta mtime failure is basically the same as "failed to read
metadata".

While we don't report in tracing log, that should be fine because dirty
reason will be reported in the other place.

### How to test and review this PR?

Should have no real user-facing behavior change.

Author: weihanglo

src/cargo/core/compiler/fingerprint/mod.rs

@@ -1167,15 +1167,12 @@ impl Fingerprint {
         // minimum mtime as it's the one we'll be comparing to inputs and
         // dependencies.
         for output in self.outputs.iter() {
-            let mtime = match paths::mtime(output) {
-                Ok(mtime) => mtime,
-
+            let Ok(mtime) = paths::mtime(output) else {
                 // This path failed to report its `mtime`. It probably doesn't
                 // exists, so leave ourselves as stale and bail out.
-                Err(e) => {
-                    debug!("failed to get mtime of {:?}: {}", output, e);
-                    return Ok(());
-                }
+                let item = StaleItem::FailedToReadMetadata(output.clone());
+                self.fs_status = FsStatus::StaleItem(item);
+                return Ok(());
             };
             assert!(mtimes.insert(output.clone(), mtime).is_none());
         }