auto merge of #19357 : michaelwoerister/rust/fix-issue-18791, r=alexcrichton

One negative side-effect of this change is that there might be quite a bit of copying strings out of the codemap, i.e. one copy for every block that gets translated, just for taking a look at the last character of the block. If this turns out to cause a performance problem then `CodeMap::span_to_snippet()` could be changed return `Option<&str>` instead of `Option<String>`.

Fixes #18791

src/test/debuginfo/multi-byte-chars.rs

@@ -0,0 +1,28 @@
+// Copyright 2013-2014 The Rust Project Developers. See the COPYRIGHT
+// file at the top-level directory of this distribution and at
+// http://rust-lang.org/COPYRIGHT.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+// ignore-android: FIXME(#10381)
+// min-lldb-version: 310
+
+// compile-flags:-g
+
+#![feature(non_ascii_idents)]
+
+// This test checks whether debuginfo generation can handle multi-byte UTF-8
+// characters at the end of a block. There's no need to do anything in the
+// debugger -- just make sure that the compiler doesn't crash.
+// See also issue #18791.
+
+struct C { θ: u8 }
+
+fn main() {
+    let x =  C { θ: 0 };
+    (|c: C| c.θ )(x);
+}