Change spans to use byte offsets instead of char offsets

src/librustc/middle/trans/debuginfo.rs

@@ -112,7 +112,7 @@ fn update_cache(cache: metadata_cache, mdtag: int, val: debug_metadata) {
 type subprogram_md = {id: ast::node_id};
 type local_var_md = {id: ast::node_id};
 type tydesc_md = {hash: uint};
-type block_md = {start: codemap::Loc<CharPos>, end: codemap::Loc<CharPos>};
+type block_md = {start: codemap::Loc, end: codemap::Loc};
 type argument_md = {id: ast::node_id};
 type retval_md = {id: ast::node_id};
 

src/libsyntax/ast_util.rs

-use codemap::{span, CharPos};
+use codemap::{span, BytePos};
 use ast::*;
 
-pure fn spanned<T>(+lo: CharPos, +hi: CharPos, +t: T) -> spanned<T> {
+pure fn spanned<T>(+lo: BytePos, +hi: BytePos, +t: T) -> spanned<T> {
     respan(mk_sp(lo, hi), move t)
 }
 
@@ -14,12 +14,12 @@
 }
 
 /* assuming that we're not in macro expansion */
-pure fn mk_sp(+lo: CharPos, +hi: CharPos) -> span {
+pure fn mk_sp(+lo: BytePos, +hi: BytePos) -> span {
     span {lo: lo, hi: hi, expn_info: None}
 }
 
 // make this a const, once the compiler supports it
-pure fn dummy_sp() -> span { return mk_sp(CharPos(0), CharPos(0)); }
+pure fn dummy_sp() -> span { return mk_sp(BytePos(0), BytePos(0)); }
 
 
 

src/libsyntax/attr.rs

@@ -6,7 +6,7 @@
 use diagnostic::span_handler;
 use ast_util::{spanned, dummy_spanned};
 use parse::comments::{doc_comment_style, strip_doc_comment_decoration};
-use codemap::CharPos;
+use codemap::BytePos;
 
 // Constructors
 export mk_name_value_item_str;
@@ -76,7 +76,7 @@ fn mk_attr(item: @ast::meta_item) -> ast::attribute {
 }
 
 fn mk_sugared_doc_attr(text: ~str,
-                       +lo: CharPos, +hi: CharPos) -> ast::attribute {
+                       +lo: BytePos, +hi: BytePos) -> ast::attribute {
     let lit = spanned(lo, hi, ast::lit_str(@text));
     let attr = {
         style: doc_comment_style(text),

src/libsyntax/codemap.rs

@@ -118,8 +118,8 @@ impl CharPos: to_bytes::IterBytes {
 }
 
 pub struct span {
-    lo: CharPos,
-    hi: CharPos,
+    lo: BytePos,
+    hi: BytePos,
     expn_info: Option<@ExpnInfo>
 }
 
@@ -141,8 +141,10 @@ impl<D: Deserializer> span: Deserializable<D> {
     }
 }
 
-pub struct Loc<A: Pos> {
-    file: @FileMap, line: uint, col: A
+// XXX col shouldn't be CharPos because col is not an absolute location in the
+// codemap, and BytePos and CharPos always represent absolute positions
+pub struct Loc {
+    file: @FileMap, line: uint, col: CharPos
 }
 
 /// An absolute offset within the CodeMap (not a relative offset within a
@@ -178,12 +180,24 @@ pub enum FileSubstr {
     pub FssExternal({filename: ~str, line: uint, col: CharPos})
 }
 
+/// Identifies an offset of a multi-byte character in a FileMap
+pub struct MultiByteChar {
+    /// The absolute offset of the character in the CodeMap
+    pos: BytePos,
+    /// The number of bytes, >=2
+    bytes: uint,
+    /// The complete number of 'extra' bytes through this character in the
+    /// FileMap
+    sum: uint
+}
+
 pub struct FileMap {
     name: FileName,
     substr: FileSubstr,
     src: @~str,
     start_pos: FilePos,
-    mut lines: ~[FilePos]
+    mut lines: ~[FilePos],
+    multibyte_chars: DVec<MultiByteChar>
 }
 
 pub impl FileMap {
@@ -194,7 +208,8 @@ pub impl FileMap {
         return FileMap {
             name: filename, substr: substr, src: src,
             start_pos: start_pos,
-            mut lines: ~[]
+            mut lines: ~[],
+            multibyte_chars: DVec()
         };
     }
 
@@ -219,6 +234,21 @@ pub fn get_line(&self, line: int) -> ~str unsafe {
         str::slice(*self.src, begin, end)
     }
 
+    pub fn record_multibyte_char(&self, pos: BytePos, bytes: uint) {
+        assert bytes >=2 && bytes <= 4;
+        let sum = if self.multibyte_chars.len() > 0 {
+            self.multibyte_chars.last().sum
+        } else {
+            0
+        };
+        let sum = sum + bytes;
+        let mbc = MultiByteChar {
+            pos: pos,
+            bytes: bytes,
+            sum: sum
+        };
+        self.multibyte_chars.push(mbc);
+    }
 }
 
 pub struct CodeMap {
@@ -254,12 +284,11 @@ pub fn mk_substr_filename(&self, sp: span) -> ~str {
                     pos.line, pos.col.to_uint());
     }
 
-    pub fn lookup_char_pos(&self, +pos: CharPos) -> Loc<CharPos> {
-        pure fn lookup(pos: FilePos) -> uint { return pos.ch.to_uint(); }
-        return self.lookup_pos(pos, lookup);
+    pub fn lookup_char_pos(&self, +pos: BytePos) -> Loc {
+        return self.lookup_pos(pos);
     }
 
-    pub fn lookup_char_pos_adj(&self, +pos: CharPos)
+    pub fn lookup_char_pos_adj(&self, +pos: BytePos)
         -> {filename: ~str, line: uint, col: CharPos, file: Option<@FileMap>}
     {
         let loc = self.lookup_char_pos(pos);
@@ -272,7 +301,7 @@ pub fn lookup_char_pos_adj(&self, +pos: CharPos)
             }
             FssInternal(sp) => {
                 self.lookup_char_pos_adj(
-                    sp.lo + (pos - loc.file.start_pos.ch))
+                    sp.lo + (pos - loc.file.start_pos.byte))
             }
             FssExternal(eloc) => {
                 {filename: /* FIXME (#2543) */ copy eloc.filename,
@@ -284,14 +313,13 @@ pub fn lookup_char_pos_adj(&self, +pos: CharPos)
     }
 
     pub fn adjust_span(&self, sp: span) -> span {
-        pure fn lookup(pos: FilePos) -> uint { return pos.ch.to_uint(); }
-        let line = self.lookup_line(sp.lo, lookup);
+        let line = self.lookup_line(sp.lo);
         match (line.fm.substr) {
             FssNone => sp,
             FssInternal(s) => {
                 self.adjust_span(span {
-                    lo: s.lo + (sp.lo - line.fm.start_pos.ch),
-                    hi: s.lo + (sp.hi - line.fm.start_pos.ch),
+                    lo: s.lo + (sp.lo - line.fm.start_pos.byte),
+                    hi: s.lo + (sp.hi - line.fm.start_pos.byte),
                     expn_info: sp.expn_info
                 })
             }
@@ -321,18 +349,6 @@ pub fn span_to_lines(&self, sp: span) -> @FileLines {
         return @FileLines {file: lo.file, lines: lines};
     }
 
-    fn lookup_byte_offset(&self, +chpos: CharPos)
-        -> {fm: @FileMap, pos: BytePos} {
-        pure fn lookup(pos: FilePos) -> uint { return pos.ch.to_uint(); }
-        let {fm, line} = self.lookup_line(chpos, lookup);
-        let line_offset = fm.lines[line].byte - fm.start_pos.byte;
-        let col = chpos - fm.lines[line].ch;
-        let col_offset = str::count_bytes(*fm.src,
-                                          line_offset.to_uint(),
-                                          col.to_uint());
-        {fm: fm, pos: line_offset + BytePos(col_offset)}
-    }
-
     pub fn span_to_snippet(&self, sp: span) -> ~str {
         let begin = self.lookup_byte_offset(sp.lo);
         let end = self.lookup_byte_offset(sp.hi);
@@ -351,15 +367,14 @@ pub fn get_filemap(&self, filename: ~str) -> @FileMap {
 }
 
 priv impl CodeMap {
-    fn lookup_line<A: Pos>(&self, pos: A, lookup: LookupFn)
-        -> {fm: @FileMap, line: uint}
-    {
+
+    fn lookup_filemap_idx(&self, +pos: BytePos) -> uint {
         let len = self.files.len();
         let mut a = 0u;
         let mut b = len;
         while b - a > 1u {
             let m = (a + b) / 2u;
-            if lookup(self.files[m].start_pos) > pos.to_uint() {
+            if self.files[m].start_pos.byte > pos {
                 b = m;
             } else {
                 a = m;
@@ -369,22 +384,40 @@ fn lookup_line<A: Pos>(&self, pos: A, lookup: LookupFn)
             fail fmt!("position %u does not resolve to a source location",
                       pos.to_uint())
         }
-        let f = self.files[a];
-        a = 0u;
-        b = vec::len(f.lines);
+
+        return a;
+    }
+
+    fn lookup_line(&self, +pos: BytePos)
+        -> {fm: @FileMap, line: uint}
+    {
+        let idx = self.lookup_filemap_idx(pos);
+        let f = self.files[idx];
+        let mut a = 0u;
+        let mut b = vec::len(f.lines);
         while b - a > 1u {
             let m = (a + b) / 2u;
-            if lookup(f.lines[m]) > pos.to_uint() { b = m; } else { a = m; }
+            if f.lines[m].byte > pos { b = m; } else { a = m; }
         }
         return {fm: f, line: a};
     }
 
-    fn lookup_pos<A: Pos Num>(&self, pos: A, lookup: LookupFn) -> Loc<A> {
-        let {fm: f, line: a} = self.lookup_line(pos, lookup);
+    fn lookup_pos(&self, +pos: BytePos) -> Loc {
+        let {fm: f, line: a} = self.lookup_line(pos);
+        let line = a + 1u; // Line numbers start at 1
+        let chpos = self.bytepos_to_local_charpos(pos);
+        let linebpos = f.lines[a].byte;
+        let linechpos = self.bytepos_to_local_charpos(linebpos);
+        debug!("codemap: byte pos %? is on the line at byte pos %?",
+               pos, linebpos);
+        debug!("codemap: char pos %? is on the line at char pos %?",
+               chpos, linechpos);
+        debug!("codemap: byte is on line: %?", line);
+        assert chpos >= linechpos;
         return Loc {
             file: f,
-            line: a + 1u,
-            col: pos - from_uint(lookup(f.lines[a]))
+            line: line,
+            col: chpos - linechpos
         };
     }
 
@@ -394,6 +427,40 @@ fn span_to_str_no_adj(&self, sp: span) -> ~str {
         return fmt!("%s:%u:%u: %u:%u", lo.file.name,
                     lo.line, lo.col.to_uint(), hi.line, hi.col.to_uint())
     }
+
+    fn lookup_byte_offset(&self, +bpos: BytePos)
+        -> {fm: @FileMap, pos: BytePos} {
+        let idx = self.lookup_filemap_idx(bpos);
+        let fm = self.files[idx];
+        let offset = bpos - fm.start_pos.byte;
+        return {fm: fm, pos: offset};
+    }
+
+    // Converts an absolute BytePos to a CharPos relative to the file it is
+    // located in
+    fn bytepos_to_local_charpos(&self, +bpos: BytePos) -> CharPos {
+        debug!("codemap: converting %? to char pos", bpos);
+        let idx = self.lookup_filemap_idx(bpos);
+        let map = self.files[idx];
+
+        // The number of extra bytes due to multibyte chars in the FileMap
+        let mut total_extra_bytes = 0;
+
+        for map.multibyte_chars.each |mbc| {
+            debug!("codemap: %?-byte char at %?", mbc.bytes, mbc.pos);
+            if mbc.pos < bpos {
+                total_extra_bytes += mbc.bytes;
+                // We should never see a byte position in the middle of a
+                // character
+                assert bpos == mbc.pos
+                    || bpos.to_uint() >= mbc.pos.to_uint() + mbc.bytes;
+            } else {
+                break;
+            }
+        }
+
+        CharPos(bpos.to_uint() - total_extra_bytes)
+    }
 }
 
 //

src/libsyntax/ext/qquote.rs

@@ -4,7 +4,7 @@
 use parse::parser::{Parser, parse_from_source_str};
 use dvec::DVec;
 use parse::token::ident_interner;
-use codemap::CharPos;
+use codemap::{CharPos, BytePos};
 
 use fold::*;
 use visit::*;
@@ -16,13 +16,13 @@
 use codemap::span;
 
 struct gather_item {
-    lo: CharPos,
-    hi: CharPos,
+    lo: BytePos,
+    hi: BytePos,
     e: @ast::expr,
     constr: ~str
 }
 
-type aq_ctxt = @{lo: CharPos, gather: DVec<gather_item>};
+type aq_ctxt = @{lo: BytePos, gather: DVec<gather_item>};
 enum fragment {
     from_expr(@ast::expr),
     from_ty(@ast::Ty)
@@ -115,7 +115,7 @@ fn mk_parse_fn(cx: ext_ctxt, sp: span) -> @ast::expr {
     fn get_fold_fn() -> ~str {~"fold_pat"}
 }
 
-fn gather_anti_quotes<N: qq_helper>(lo: CharPos, node: N) -> aq_ctxt
+fn gather_anti_quotes<N: qq_helper>(lo: BytePos, node: N) -> aq_ctxt
 {
     let v = @{visit_expr: |node, &&cx, v| visit_aq(node, ~"from_expr", cx, v),
               visit_ty: |node, &&cx, v| visit_aq(node, ~"from_ty", cx, v),
@@ -227,7 +227,7 @@ fn finish<T: qq_helper>
     let mut str2 = ~"";
     enum state {active, skip(uint), blank};
     let mut state = active;
-    let mut i = CharPos(0u);
+    let mut i = BytePos(0u);
     let mut j = 0u;
     let g_len = cx.gather.len();
     for str::chars_each(*str) |ch| {
@@ -244,7 +244,7 @@ fn finish<T: qq_helper>
           blank if is_space(ch) => str::push_char(&mut str2, ch),
           blank => str::push_char(&mut str2, ' ')
         }
-        i += CharPos(1u);
+        i += BytePos(1u);
         if (j < g_len && i == cx.gather[j].hi) {
             assert ch == ')';
             state = active;

src/libsyntax/ext/tt/macro_parser.rs

@@ -11,7 +11,7 @@
 use ast::{matcher, match_tok, match_seq, match_nonterminal, ident};
 use ast_util::mk_sp;
 use std::map::HashMap;
-use codemap::CharPos;
+use codemap::BytePos;
 
 /* This is an Earley-like parser, without support for in-grammar nonterminals,
 only by calling out to the main rust parser for named nonterminals (which it
@@ -103,7 +103,7 @@ fn is_some(&&mpu: matcher_pos_up) -> bool {
     mut up: matcher_pos_up, // mutable for swapping only
     matches: ~[DVec<@named_match>],
     match_lo: uint, match_hi: uint,
-    sp_lo: CharPos,
+    sp_lo: BytePos,
 };
 
 fn copy_up(&& mpu: matcher_pos_up) -> matcher_pos {
@@ -123,7 +123,7 @@ fn count_names(ms: &[matcher]) -> uint {
 }
 
 #[allow(non_implicitly_copyable_typarams)]
-fn initial_matcher_pos(ms: ~[matcher], sep: Option<Token>, lo: CharPos)
+fn initial_matcher_pos(ms: ~[matcher], sep: Option<Token>, lo: BytePos)
     -> matcher_pos {
     let mut match_idx_hi = 0u;
     for ms.each() |elt| {

src/libsyntax/parse/attr.rs

@@ -14,7 +14,7 @@ fn parse_outer_attrs_or_ext(first_item_attrs: ~[ast::attribute])
         -> attr_or_ext;
     fn parse_outer_attributes() -> ~[ast::attribute];
     fn parse_attribute(style: ast::attr_style) -> ast::attribute;
-    fn parse_attribute_naked(style: ast::attr_style, lo: CharPos) ->
+    fn parse_attribute_naked(style: ast::attr_style, lo: BytePos) ->
         ast::attribute;
     fn parse_inner_attrs_and_next() ->
         {inner: ~[ast::attribute], next: ~[ast::attribute]};
@@ -85,7 +85,7 @@ fn parse_attribute(style: ast::attr_style) -> ast::attribute {
         return self.parse_attribute_naked(style, lo);
     }
 
-    fn parse_attribute_naked(style: ast::attr_style, lo: CharPos) ->
+    fn parse_attribute_naked(style: ast::attr_style, lo: BytePos) ->
         ast::attribute {
         self.expect(token::LBRACKET);
         let meta_item = self.parse_meta_item();

src/libsyntax/parse/comments.rs

@@ -28,7 +28,7 @@ impl cmnt_style : cmp::Eq {
     }
 }
 
-type cmnt = {style: cmnt_style, lines: ~[~str], pos: CharPos};
+type cmnt = {style: cmnt_style, lines: ~[~str], pos: BytePos};
 
 fn is_doc_comment(s: ~str) -> bool {
     s.starts_with(~"///") ||
@@ -131,7 +131,7 @@ fn consume_non_eol_whitespace(rdr: string_reader) {
 fn push_blank_line_comment(rdr: string_reader, comments: &mut ~[cmnt]) {
     debug!(">>> blank-line comment");
     let v: ~[~str] = ~[];
-    comments.push({style: blank_line, lines: v, pos: rdr.last_pos.ch});
+    comments.push({style: blank_line, lines: v, pos: rdr.last_pos.byte});
 }
 
 fn consume_whitespace_counting_blank_lines(rdr: string_reader,
@@ -148,7 +148,7 @@ fn consume_whitespace_counting_blank_lines(rdr: string_reader,
 fn read_shebang_comment(rdr: string_reader, code_to_the_left: bool,
                                             comments: &mut ~[cmnt]) {
     debug!(">>> shebang comment");
-    let p = rdr.last_pos.ch;
+    let p = rdr.last_pos.byte;
     debug!("<<< shebang comment");
     comments.push({
         style: if code_to_the_left { trailing } else { isolated },
@@ -160,7 +160,7 @@ fn read_shebang_comment(rdr: string_reader, code_to_the_left: bool,
 fn read_line_comments(rdr: string_reader, code_to_the_left: bool,
                                           comments: &mut ~[cmnt]) {
     debug!(">>> line comments");
-    let p = rdr.last_pos.ch;
+    let p = rdr.last_pos.byte;
     let mut lines: ~[~str] = ~[];
     while rdr.curr == '/' && nextch(rdr) == '/' {
         let line = read_one_line_comment(rdr);
@@ -209,7 +209,7 @@ fn trim_whitespace_prefix_and_push_line(lines: &mut ~[~str],
 fn read_block_comment(rdr: string_reader, code_to_the_left: bool,
                                           comments: &mut ~[cmnt]) {
     debug!(">>> block comment");
-    let p = rdr.last_pos.ch;
+    let p = rdr.last_pos.byte;
     let mut lines: ~[~str] = ~[];
     let mut col: CharPos = rdr.col;
     bump(rdr);
@@ -284,7 +284,7 @@ fn consume_comment(rdr: string_reader, code_to_the_left: bool,
     debug!("<<< consume comment");
 }
 
-type lit = {lit: ~str, pos: CharPos};
+type lit = {lit: ~str, pos: BytePos};
 
 fn gather_comments_and_literals(span_diagnostic: diagnostic::span_handler,
                                 path: ~str,

src/libsyntax/parse/common.rs

@@ -205,7 +205,7 @@ fn expect_gt() {
         if self.token == token::GT {
             self.bump();
         } else if self.token == token::BINOP(token::SHR) {
-            self.swap(token::GT, self.span.lo + CharPos(1u), self.span.hi);
+            self.swap(token::GT, self.span.lo + BytePos(1u), self.span.hi);
         } else {
             let mut s: ~str = ~"expected `";
             s += token_to_str(self.reader, token::GT);

src/libsyntax/parse/lexer.rs

@@ -23,7 +23,7 @@ trait reader {
     src: @~str,
     // The absolute offset within the codemap of the next character to read
     mut pos: FilePos,
-    // The absolute offset within the codemap of the last character to be read (curr)
+    // The absolute offset within the codemap of the last character read(curr)
     mut last_pos: FilePos,
     // The column of the next character to read
     mut col: CharPos,
@@ -123,9 +123,9 @@ fn string_advance_token(&&r: string_reader) {
     if is_eof(r) {
         r.peek_tok = token::EOF;
     } else {
-        let start_chpos = r.last_pos.ch;
+        let start_bytepos = r.last_pos.byte;
         r.peek_tok = next_token_inner(r);
-        r.peek_span = ast_util::mk_sp(start_chpos, r.last_pos.ch);
+        r.peek_span = ast_util::mk_sp(start_bytepos, r.last_pos.byte);
     };
 
 }
@@ -158,6 +158,11 @@ fn bump(rdr: string_reader) {
             rdr.filemap.next_line(rdr.last_pos);
             rdr.col = CharPos(0u);
         }
+
+        if byte_offset_diff > 1 {
+            rdr.filemap.record_multibyte_char(
+                BytePos(current_byte_offset), byte_offset_diff);
+        }
     } else {
         // XXX: What does this accomplish?
         if (rdr.curr != -1 as char) {
@@ -233,7 +238,7 @@ fn consume_any_line_comment(rdr: string_reader)
             bump(rdr);
             // line comments starting with "///" or "//!" are doc-comments
             if rdr.curr == '/' || rdr.curr == '!' {
-                let start_chpos = rdr.pos.ch - CharPos(2u);
+                let start_bpos = rdr.pos.byte - BytePos(2u);
                 let mut acc = ~"//";
                 while rdr.curr != '\n' && !is_eof(rdr) {
                     str::push_char(&mut acc, rdr.curr);
@@ -241,7 +246,7 @@ fn consume_any_line_comment(rdr: string_reader)
                 }
                 return Some({
                     tok: token::DOC_COMMENT(rdr.interner.intern(@acc)),
-                    sp: ast_util::mk_sp(start_chpos, rdr.pos.ch)
+                    sp: ast_util::mk_sp(start_bpos, rdr.pos.byte)
                 });
             } else {
                 while rdr.curr != '\n' && !is_eof(rdr) { bump(rdr); }
@@ -256,7 +261,7 @@ fn consume_any_line_comment(rdr: string_reader)
         if nextch(rdr) == '!' {
             let cmap = @CodeMap::new();
             (*cmap).files.push(rdr.filemap);
-            let loc = cmap.lookup_char_pos_adj(rdr.last_pos.ch);
+            let loc = cmap.lookup_char_pos_adj(rdr.last_pos.byte);
             if loc.line == 1u && loc.col == CharPos(0u) {
                 while rdr.curr != '\n' && !is_eof(rdr) { bump(rdr); }
                 return consume_whitespace_and_comments(rdr);
@@ -272,7 +277,7 @@ fn consume_block_comment(rdr: string_reader)
 
     // block comments starting with "/**" or "/*!" are doc-comments
     if rdr.curr == '*' || rdr.curr == '!' {
-        let start_chpos = rdr.pos.ch - CharPos(2u);
+        let start_bpos = rdr.pos.byte - BytePos(2u);
         let mut acc = ~"/*";
         while !(rdr.curr == '*' && nextch(rdr) == '/') && !is_eof(rdr) {
             str::push_char(&mut acc, rdr.curr);
@@ -286,7 +291,7 @@ fn consume_block_comment(rdr: string_reader)
             bump(rdr);
             return Some({
                 tok: token::DOC_COMMENT(rdr.interner.intern(@acc)),
-                sp: ast_util::mk_sp(start_chpos, rdr.pos.ch)
+                sp: ast_util::mk_sp(start_bpos, rdr.pos.byte)
             });
         }
     } else {

src/libsyntax/parse/parser.rs

@@ -5,7 +5,7 @@
 use std::map::HashMap;
 use token::{can_begin_expr, is_ident, is_ident_or_path, is_plain_ident,
             INTERPOLATED, special_idents};
-use codemap::{span,FssNone, CharPos};
+use codemap::{span,FssNone, BytePos};
 use util::interner::Interner;
 use ast_util::{spanned, respan, mk_sp, ident_to_path, operator_prec};
 use lexer::reader;
@@ -244,7 +244,7 @@ fn bump() {
         self.token = next.tok;
         self.span = next.sp;
     }
-    fn swap(next: token::Token, +lo: CharPos, +hi: CharPos) {
+    fn swap(next: token::Token, +lo: BytePos, +hi: BytePos) {
         self.token = next;
         self.span = mk_sp(lo, hi);
     }
@@ -904,12 +904,12 @@ fn parse_field(sep: token::Token) -> field {
         return spanned(lo, e.span.hi, {mutbl: m, ident: i, expr: e});
     }
 
-    fn mk_expr(+lo: CharPos, +hi: CharPos, +node: expr_) -> @expr {
+    fn mk_expr(+lo: BytePos, +hi: BytePos, +node: expr_) -> @expr {
         return @{id: self.get_id(), callee_id: self.get_id(),
               node: node, span: mk_sp(lo, hi)};
     }
 
-    fn mk_mac_expr(+lo: CharPos, +hi: CharPos, m: mac_) -> @expr {
+    fn mk_mac_expr(+lo: BytePos, +hi: BytePos, m: mac_) -> @expr {
         return @{id: self.get_id(),
               callee_id: self.get_id(),
               node: expr_mac({node: m, span: mk_sp(lo, hi)}),
@@ -1134,7 +1134,7 @@ fn parse_bottom_expr() -> @expr {
         return self.mk_expr(lo, hi, ex);
     }
 
-    fn parse_block_expr(lo: CharPos, blk_mode: blk_check_mode) -> @expr {
+    fn parse_block_expr(lo: BytePos, blk_mode: blk_check_mode) -> @expr {
         self.expect(token::LBRACE);
         let blk = self.parse_block_tail(lo, blk_mode);
         return self.mk_expr(blk.span.lo, blk.span.hi, expr_block(blk));
@@ -1146,7 +1146,7 @@ fn parse_syntax_ext() -> @expr {
         return self.parse_syntax_ext_naked(lo);
     }
 
-    fn parse_syntax_ext_naked(lo: CharPos) -> @expr {
+    fn parse_syntax_ext_naked(lo: BytePos) -> @expr {
         match self.token {
           token::IDENT(_, _) => (),
           _ => self.fatal(~"expected a syntax expander name")
@@ -2279,11 +2279,11 @@ fn parse_block_no_value() -> blk {
     // I guess that also means "already parsed the 'impure'" if
     // necessary, and this should take a qualifier.
     // some blocks start with "#{"...
-    fn parse_block_tail(lo: CharPos, s: blk_check_mode) -> blk {
+    fn parse_block_tail(lo: BytePos, s: blk_check_mode) -> blk {
         self.parse_block_tail_(lo, s, ~[])
     }
 
-    fn parse_block_tail_(lo: CharPos, s: blk_check_mode,
+    fn parse_block_tail_(lo: BytePos, s: blk_check_mode,
                          +first_item_attrs: ~[attribute]) -> blk {
         let mut stmts = ~[];
         let mut expr = None;
@@ -2581,7 +2581,7 @@ fn parse_fn_header() -> {ident: ident, tps: ~[ty_param]} {
         return {ident: id, tps: ty_params};
     }
 
-    fn mk_item(+lo: CharPos, +hi: CharPos, +ident: ident,
+    fn mk_item(+lo: BytePos, +hi: BytePos, +ident: ident,
                +node: item_, vis: visibility,
                +attrs: ~[attribute]) -> @item {
         return @{ident: ident,
@@ -3037,7 +3037,7 @@ fn parse_foreign_mod_items(sort: ast::foreign_mod_sort,
             items: items};
     }
 
-    fn parse_item_foreign_mod(lo: CharPos,
+    fn parse_item_foreign_mod(lo: BytePos,
                               visibility: visibility,
                               attrs: ~[attribute],
                               items_allowed: bool)
@@ -3092,7 +3092,7 @@ fn parse_item_foreign_mod(lo: CharPos,
         });
     }
 
-    fn parse_type_decl() -> {lo: CharPos, ident: ident} {
+    fn parse_type_decl() -> {lo: BytePos, ident: ident} {
         let lo = self.last_span.lo;
         let id = self.parse_ident();
         return {lo: lo, ident: id};

src/libsyntax/print/pprust.rs

 use parse::{comments, lexer, token};
-use codemap::{CodeMap, CharPos};
+use codemap::{CodeMap, BytePos};
 use pp::{break_offset, word, printer, space, zerobreak, hardbreak, breaks};
 use pp::{consistent, inconsistent, eof};
 use ast::{required, provided};
@@ -1898,7 +1898,7 @@ fn print_ty_fn(s: ps,
 }
 
 fn maybe_print_trailing_comment(s: ps, span: codemap::span,
-                                next_pos: Option<CharPos>) {
+                                next_pos: Option<BytePos>) {
     let mut cm;
     match s.cm { Some(ccm) => cm = ccm, _ => return }
     match next_comment(s) {
@@ -1906,7 +1906,7 @@ fn maybe_print_trailing_comment(s: ps, span: codemap::span,
         if cmnt.style != comments::trailing { return; }
         let span_line = cm.lookup_char_pos(span.hi);
         let comment_line = cm.lookup_char_pos(cmnt.pos);
-        let mut next = cmnt.pos + CharPos(1u);
+        let mut next = cmnt.pos + BytePos(1u);
         match next_pos { None => (), Some(p) => next = p }
         if span.hi < cmnt.pos && cmnt.pos < next &&
                span_line.line == comment_line.line {
@@ -1981,7 +1981,7 @@ fn lit_to_str(l: @ast::lit) -> ~str {
     return to_str(l, print_literal, parse::token::mk_fake_ident_interner());
 }
 
-fn next_lit(s: ps, pos: CharPos) -> Option<comments::lit> {
+fn next_lit(s: ps, pos: BytePos) -> Option<comments::lit> {
     match s.literals {
       Some(lits) => {
         while s.cur_lit < vec::len(lits) {
@@ -1996,7 +1996,7 @@ fn next_lit(s: ps, pos: CharPos) -> Option<comments::lit> {
     }
 }
 
-fn maybe_print_comment(s: ps, pos: CharPos) {
+fn maybe_print_comment(s: ps, pos: BytePos) {
     loop {
         match next_comment(s) {
           Some(cmnt) => {