Rollup merge of rust-lang#48522 - etaoins:fix-find-width-of-character-at-span-bounds-check, r=estebank

Fix find_width_of_character_at_span bounds check

Commit 0bd9667 added bounds checking of our current target byte position to prevent infinite loops. Unfortunately it was comparing the file-relative `target` versus the global `file_start_pos` and `file_end_pos`.

The result is failing to detect multibyte characters unless their file-relative offset fit within their global offset. This causes other parts of the compiler to generate spans pointing to the middle of a
multibyte character which will ultimately panic in `bytepos_to_file_charpos`.

Fix by comparing the `target` to the total file size when moving forward and doing checked subtraction when moving backwards. This should preserve the intent of the bounds check while removing the offset confusion.

cc @davidtwco

Fixes rust-lang#48508

Author: Manishearth

src/libsyntax/codemap.rs

@@ -705,17 +705,20 @@ impl CodeMap {
         };
         debug!("find_width_of_character_at_span: snippet=`{:?}`", snippet);
 
-        let file_start_pos = local_begin.fm.start_pos.to_usize();
-        let file_end_pos = local_begin.fm.end_pos.to_usize();
-        debug!("find_width_of_character_at_span: file_start_pos=`{:?}` file_end_pos=`{:?}`",
-               file_start_pos, file_end_pos);
-
         let mut target = if forwards { end_index + 1 } else { end_index - 1 };
         debug!("find_width_of_character_at_span: initial target=`{:?}`", target);
 
-        while !snippet.is_char_boundary(target - start_index)
-                && target >= file_start_pos && target <= file_end_pos {
-            target = if forwards { target + 1 } else { target - 1 };
+        while !snippet.is_char_boundary(target - start_index) && target < source_len {
+            target = if forwards {
+                target + 1
+            } else {
+                match target.checked_sub(1) {
+                    Some(target) => target,
+                    None => {
+                        break;
+                    }
+                }
+            };
             debug!("find_width_of_character_at_span: target=`{:?}`", target);
         }
         debug!("find_width_of_character_at_span: final target=`{:?}`", target);

src/test/run-pass/issue-48508-aux.rs

@@ -0,0 +1,16 @@
+// Copyright 2018 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-test Not a test. Used by issue-48508.rs
+
+pub fn other() -> f64 {
+    let µ = 1.0;
+    µ
+}

src/test/run-pass/issue-48508.rs

@@ -0,0 +1,28 @@
+// Copyright 2018 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.
+
+// Regression test for issue #48508:
+//
+// Confusion between global and local file offsets caused incorrect handling of multibyte character
+// spans when compiling multiple files. One visible effect was an ICE generating debug information
+// when a multibyte character is at the end of a scope. The problematic code is actually in
+// issue-48508-aux.rs
+
+// compile-flags:-g
+// ignore-pretty issue #37195
+
+#![feature(non_ascii_idents)]
+
+#[path = "issue-48508-aux.rs"]
+mod other_file;
+
+fn main() {
+    other_file::other();
+}