syntax: store char literals/tokens as chars rather than u32s.

Clearly storing them as `char` is semantically nicer, but this also
fixes a bug whereby `quote_expr!(cx, 'a')` wasn't working, because the
code created by quotation was not matching the actual AST definitions.

Author: huonw

src/librustdoc/clean.rs

@@ -26,7 +26,6 @@ use rustc::metadata::decoder;
 
 use std::local_data;
 use std::strbuf::StrBuf;
-use std;
 
 use core;
 use doctree;
@@ -1246,7 +1245,7 @@ fn lit_to_str(lit: &ast::Lit) -> ~str {
     match lit.node {
         ast::LitStr(ref st, _) => st.get().to_owned(),
         ast::LitBinary(ref data) => format!("{:?}", data.as_slice()),
-        ast::LitChar(c) => "'".to_owned() + std::char::from_u32(c).unwrap().to_str() + "'",
+        ast::LitChar(c) => format!("'{}'", c),
         ast::LitInt(i, _t) => i.to_str(),
         ast::LitUint(u, _t) => u.to_str(),
         ast::LitIntUnsuffixed(i) => i.to_str(),

src/libsyntax/ast.rs

@@ -653,7 +653,7 @@ pub type Lit = Spanned<Lit_>;
 pub enum Lit_ {
     LitStr(InternedString, StrStyle),
     LitBinary(Rc<Vec<u8> >),
-    LitChar(u32),
+    LitChar(char),
     LitInt(i64, IntTy),
     LitUint(u64, UintTy),
     LitIntUnsuffixed(i64),

src/libsyntax/ext/bytes.rs

@@ -16,8 +16,6 @@ use ext::base::*;
 use ext::base;
 use ext::build::AstBuilder;
 
-use std::char;
-
 pub fn expand_syntax_ext(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree]) -> ~base::MacResult {
     // Gather all argument expressions
     let exprs = match get_exprs_from_tts(cx, sp, tts) {
@@ -59,7 +57,7 @@ pub fn expand_syntax_ext(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree]) ->
 
                 // char literal, push to vector expression
                 ast::LitChar(v) => {
-                    if char::from_u32(v).unwrap().is_ascii() {
+                    if v.is_ascii() {
                         bytes.push(cx.expr_u8(expr.span, v as u8));
                     } else {
                         cx.span_err(expr.span, "non-ascii char literal in bytes!")

src/libsyntax/ext/concat.rs

@@ -14,7 +14,6 @@ use ext::base;
 use ext::build::AstBuilder;
 use parse::token;
 
-use std::char;
 use std::strbuf::StrBuf;
 
 pub fn expand_syntax_ext(cx: &mut base::ExtCtxt,
@@ -35,7 +34,7 @@ pub fn expand_syntax_ext(cx: &mut base::ExtCtxt,
                         accumulator.push_str(s.get());
                     }
                     ast::LitChar(c) => {
-                        accumulator.push_char(char::from_u32(c).unwrap());
+                        accumulator.push_char(c);
                     }
                     ast::LitInt(i, _) | ast::LitIntUnsuffixed(i) => {
                         accumulator.push_str(format!("{}", i));

src/libsyntax/ext/format.rs

@@ -561,7 +561,7 @@ impl<'a, 'b> Context<'a, 'b> {
 
                 // Translate the format
                 let fill = match arg.format.fill { Some(c) => c, None => ' ' };
-                let fill = self.ecx.expr_lit(sp, ast::LitChar(fill as u32));
+                let fill = self.ecx.expr_lit(sp, ast::LitChar(fill));
                 let align = match arg.format.align {
                     parse::AlignLeft => {
                         self.ecx.path_global(sp, self.parsepath("AlignLeft"))

src/libsyntax/parse/lexer.rs

@@ -874,7 +874,7 @@ fn next_token_inner(rdr: &mut StringReader) -> token::Token {
                                "unterminated character constant".to_owned());
         }
         bump(rdr); // advance curr past token
-        return token::LIT_CHAR(c2 as u32);
+        return token::LIT_CHAR(c2);
       }
       '"' => {
         let mut accum_str = StrBuf::new();
@@ -1097,17 +1097,17 @@ mod test {
 
     #[test] fn character_a() {
         assert_eq!(setup(&mk_sh(), "'a'".to_owned()).next_token().tok,
-                   token::LIT_CHAR('a' as u32));
+                   token::LIT_CHAR('a'));
     }
 
     #[test] fn character_space() {
         assert_eq!(setup(&mk_sh(), "' '".to_owned()).next_token().tok,
-                   token::LIT_CHAR(' ' as u32));
+                   token::LIT_CHAR(' '));
     }
 
     #[test] fn character_escaped() {
         assert_eq!(setup(&mk_sh(), "'\\n'".to_owned()).next_token().tok,
-                   token::LIT_CHAR('\n' as u32));
+                   token::LIT_CHAR('\n'));
     }
 
     #[test] fn lifetime_name() {
@@ -1128,7 +1128,7 @@ mod test {
 
     #[test] fn nested_block_comments() {
         assert_eq!(setup(&mk_sh(), "/* /* */ */'a'".to_owned()).next_token().tok,
-                   token::LIT_CHAR('a' as u32));
+                   token::LIT_CHAR('a'));
     }
 
 }

src/libsyntax/parse/token.rs

@@ -18,7 +18,6 @@ use util::interner;
 
 use serialize::{Decodable, Decoder, Encodable, Encoder};
 use std::cast;
-use std::char;
 use std::fmt;
 use std::local_data;
 use std::path::BytesContainer;
@@ -81,7 +80,7 @@ pub enum Token {
     DOLLAR,
 
     /* Literals */
-    LIT_CHAR(u32),
+    LIT_CHAR(char),
     LIT_INT(i64, ast::IntTy),
     LIT_UINT(u64, ast::UintTy),
     LIT_INT_UNSUFFIXED(i64),
@@ -195,7 +194,7 @@ pub fn to_str(t: &Token) -> ~str {
       /* Literals */
       LIT_CHAR(c) => {
           let mut res = StrBuf::from_str("'");
-          char::from_u32(c).unwrap().escape_default(|c| {
+          c.escape_default(|c| {
               res.push_char(c);
           });
           res.push_char('\'');

src/libsyntax/print/pprust.rs

@@ -26,7 +26,6 @@ use print::pp::{Breaks, Consistent, Inconsistent, eof};
 use print::pp;
 
 use std::cast;
-use std::char;
 use std::io::{IoResult, MemWriter};
 use std::io;
 use std::rc::Rc;
@@ -2196,7 +2195,7 @@ impl<'a> State<'a> {
             ast::LitStr(ref st, style) => self.print_string(st.get(), style),
             ast::LitChar(ch) => {
                 let mut res = StrBuf::from_str("'");
-                char::from_u32(ch).unwrap().escape_default(|c| res.push_char(c));
+                ch.escape_default(|c| res.push_char(c));
                 res.push_char('\'');
                 word(&mut self.s, res.into_owned())
             }