libsyntax: librustdoc: ignore utf-8 BOM in .rs files

Closes #12974

1 files changed, 10 insertions, 1 deletions

diff --git a/src/libsyntax/codemap.rs b/src/libsyntax/codemap.rs
index 4bfd5391a8f..d93b5803eac 100644
--- a/src/libsyntax/codemap.rs
+++ b/src/libsyntax/codemap.rs
@@ -271,13 +271,22 @@ impl CodeMap {
         }
     }
 
-    pub fn new_filemap(&self, filename: FileName, mut src: ~str) -> Rc<FileMap> {
+    pub fn new_filemap(&self, filename: FileName, src: ~str) -> Rc<FileMap> {
         let mut files = self.files.borrow_mut();
         let start_pos = match files.get().last() {
             None => 0,
             Some(last) => last.deref().start_pos.to_uint() + last.deref().src.len(),
         };
 
+        // Remove utf-8 BOM if any.
+        // FIXME #12884: no efficient/safe way to remove from the start of a string
+        // and reuse the allocation.
+        let mut src = if src.starts_with("\ufeff") {
+            src.as_slice().slice_from(3).into_owned()
+        } else {
+            src
+        };
+
         // Append '\n' in case it's not already there.
         // This is a workaround to prevent CodeMap.lookup_filemap_idx from accidentally
         // overflowing into the next filemap in case the last byte of span is also the last