Rollup merge of rust-lang#134366 - harrisonkaiser:no-break-space, r=davidtwco

Fix logical error with what text is considered whitespace.

There appears to be a logical issue around what counts as leading white-space. There is code which does a subtraction assuming that no errors will be reported inside the leading whitespace. However we compute the length of that whitespace with std::char::is_whitespace and not rustc_lexer::is_whitespace. The former will include a no-break space while later will excluded it. We can only safely make the assumption that no errors will be reported  in whitespace if it is all "Rust Standard" whitespace. Indeed an error does occur in unicode whitespace if it contains a no-break space. In that case the subtraction will cause a ICE (for a compiler in debug mode) as described in rust-lang#132918.

Author: jieyouxu

Cargo.lock

@@ -3722,6 +3722,7 @@ dependencies = [
  "rustc_fluent_macro",
  "rustc_hir",
  "rustc_index",
+ "rustc_lexer",
  "rustc_lint_defs",
  "rustc_macros",
  "rustc_serialize",

compiler/rustc_errors/Cargo.toml

@@ -16,6 +16,7 @@ rustc_error_messages = { path = "../rustc_error_messages" }
 rustc_fluent_macro = { path = "../rustc_fluent_macro" }
 rustc_hir = { path = "../rustc_hir" }
 rustc_index = { path = "../rustc_index" }
+rustc_lexer = { path = "../rustc_lexer" }
 rustc_lint_defs = { path = "../rustc_lint_defs" }
 rustc_macros = { path = "../rustc_macros" }
 rustc_serialize = { path = "../rustc_serialize" }

compiler/rustc_errors/src/emitter.rs

@@ -19,6 +19,7 @@ use derive_setters::Setters;
 use rustc_data_structures::fx::{FxHashMap, FxIndexMap, FxIndexSet};
 use rustc_data_structures::sync::{DynSend, IntoDynSyncSend, Lrc};
 use rustc_error_messages::{FluentArgs, SpanLabel};
+use rustc_lexer;
 use rustc_lint_defs::pluralize;
 use rustc_span::hygiene::{ExpnKind, MacroKind};
 use rustc_span::source_map::SourceMap;
@@ -1698,9 +1699,14 @@ impl HumanEmitter {
                     if let Some(source_string) =
                         line.line_index.checked_sub(1).and_then(|l| file.get_line(l))
                     {
+                        // Whitespace can only be removed (aka considered leading)
+                        // if the lexer considers it whitespace.
+                        // non-rustc_lexer::is_whitespace() chars are reported as an
+                        // error (ex. no-break-spaces \u{a0}), and thus can't be considered
+                        // for removal during error reporting.
                         let leading_whitespace = source_string
                             .chars()
-                            .take_while(|c| c.is_whitespace())
+                            .take_while(|c| rustc_lexer::is_whitespace(*c))
                             .map(|c| {
                                 match c {
                                     // Tabs are displayed as 4 spaces
@@ -1709,7 +1715,7 @@ impl HumanEmitter {
                                 }
                             })
                             .sum();
-                        if source_string.chars().any(|c| !c.is_whitespace()) {
+                        if source_string.chars().any(|c| !rustc_lexer::is_whitespace(c)) {
                             whitespace_margin = min(whitespace_margin, leading_whitespace);
                         }
                     }

tests/ui/errors/emitter-overflow-bad-whitespace.rs

@@ -0,0 +1,11 @@
+// Invalid whitespace (not listed here: https://doc.rust-lang.org/reference/whitespace.html
+// e.g. \u{a0}) before any other syntax on the line should not cause any integer overflow
+// in the emitter, even when the terminal width causes the line to be truncated.
+//
+// issue #132918
+
+//@ check-fail
+//@ needs-rustc-debug-assertions
+//@ compile-flags: --diagnostic-width=1
+                                         fn main() {              return;              }
+//~^ ERROR unknown start of token: \u{a0}

tests/ui/errors/emitter-overflow-bad-whitespace.stderr

@@ -0,0 +1,13 @@
+error: unknown start of token: \u{a0}
+  --> $DIR/emitter-overflow-bad-whitespace.rs:10:1
+   |
+LL |     ...
+   | ^
+   |
+help: Unicode character ' ' (No-Break Space) looks like ' ' (Space), but it is not
+   |
+LL |                                          fn main() {              return;              }
+   | +
+
+error: aborting due to 1 previous error
+