bjoernager/rust
bjoernager/rust
1
Fork 0
Code Activity

auto merge of #15339 : cmr/rust/rewrite-lexer2, r=huonw

Browse source 
Mostly minor things that rebasing is becoming painful.
This commit is contained in:
bors 2014-07-09 08:46:35 +00:00
commit ad3eda123a
63 changed files with 2300 additions and 2066 deletions
Download patch file Download diff file Expand all files Collapse all files

7
src/libcore/str.rs
View file

@ -1764,7 +1764,9 @@ impl<'a> StrSlice<'a> for &'a str {
#[inline]
fn slice(&self, begin: uint, end: uint) -> &'a str {
assert!(self.is_char_boundary(begin) && self.is_char_boundary(end));
assert!(self.is_char_boundary(begin) && self.is_char_boundary(end),
"index {} and/or {} in `{}` do not lie on character boundary", begin,
end, *self);
unsafe { raw::slice_bytes(*self, begin, end) }
}
@ -1775,7 +1777,8 @@ impl<'a> StrSlice<'a> for &'a str {
#[inline]
fn slice_to(&self, end: uint) -> &'a str {
assert!(self.is_char_boundary(end));
assert!(self.is_char_boundary(end), "index {} in `{}` does not lie on \
a character boundary", end, *self);
unsafe { raw::slice_bytes(*self, 0, end) }
}

2
src/librustc/lint/builtin.rs
View file

@ -1114,7 +1114,7 @@ impl UnusedMut {
match mode {
ast::BindByValue(ast::MutMutable) => {
if !token::get_ident(ident).get().starts_with("_") {
mutables.insert_or_update_with(ident.name as uint,
mutables.insert_or_update_with(ident.name.uint(),
vec!(id), |_, old| { old.push(id); });
}
}

2
src/librustc/metadata/decoder.rs
View file

@ -323,7 +323,7 @@ fn item_name(intr: &IdentInterner, item: ebml::Doc) -> ast::Ident {
let string = name.as_str_slice();
match intr.find_equiv(&string) {
None => token::str_to_ident(string),
Some(val) => ast::Ident::new(val as ast::Name),
Some(val) => ast::Ident::new(val),
}
}

3
src/librustc/middle/astencode.rs
View file

@ -1523,7 +1523,7 @@ fn test_basic() {
fn foo() {}
));
}
/* NOTE: When there's a snapshot, update this (yay quasiquoter!)
#[test]
fn test_smalltalk() {
let cx = mk_ctxt();
@ -1531,6 +1531,7 @@ fn test_smalltalk() {
fn foo() -> int { 3 + 4 } // first smalltalk program ever executed.
));
}
*/
#[test]
fn test_more() {

1
src/librustc/middle/trans/consts.rs
View file

@ -42,6 +42,7 @@ use syntax::{ast, ast_util};
pub fn const_lit(cx: &CrateContext, e: &ast::Expr, lit: ast::Lit)
-> ValueRef {
let _icx = push_ctxt("trans_lit");
debug!("const_lit: {}", lit);
match lit.node {
ast::LitByte(b) => C_integral(Type::uint_from_ty(cx, ast::TyU8), b as u64, false),
ast::LitChar(i) => C_integral(Type::char(cx), i as u64, false),

42
src/librustdoc/html/highlight.rs
View file

@ -18,7 +18,6 @@ use std::io;
use syntax::parse;
use syntax::parse::lexer;
use syntax::codemap::{BytePos, Span};
use html::escape::Escape;
@ -59,38 +58,30 @@ fn doit(sess: &parse::ParseSess, mut lexer: lexer::StringReader,
None => {}
}
try!(write!(out, "class='rust {}'>\n", class.unwrap_or("")));
let mut last = BytePos(0);
let mut is_attribute = false;
let mut is_macro = false;
let mut is_macro_nonterminal = false;
loop {
let next = lexer.next_token();
let test = if next.tok == t::EOF {lexer.pos} else {next.sp.lo};
// The lexer consumes all whitespace and non-doc-comments when iterating
// between tokens. If this token isn't directly adjacent to our last
// token, then we need to emit the whitespace/comment.
//
// If the gap has any '/' characters then we consider the whole thing a
// comment. This will classify some whitespace as a comment, but that
// doesn't matter too much for syntax highlighting purposes.
if test > last {
let snip = sess.span_diagnostic.cm.span_to_snippet(Span {
lo: last,
hi: test,
expn_info: None,
}).unwrap();
if snip.as_slice().contains("/") {
try!(write!(out, "<span class='comment'>{}</span>",
Escape(snip.as_slice())));
} else {
try!(write!(out, "{}", Escape(snip.as_slice())));
}
}
last = next.sp.hi;
let snip = |sp| sess.span_diagnostic.cm.span_to_snippet(sp).unwrap();
if next.tok == t::EOF { break }
let klass = match next.tok {
t::WS => {
try!(write!(out, "{}", Escape(snip(next.sp).as_slice())));
continue
},
t::COMMENT => {
try!(write!(out, "<span class='comment'>{}</span>",
Escape(snip(next.sp).as_slice())));
continue
},
t::SHEBANG(s) => {
try!(write!(out, "{}", Escape(s.as_str())));
continue
},
// If this '&' token is directly adjacent to another token, assume
// that it's the address-of operator instead of the and-operator.
// This allows us to give all pointers their own class (`Box` and
@ -144,8 +135,7 @@ fn doit(sess: &parse::ParseSess, mut lexer: lexer::StringReader,
t::LIT_CHAR(..) | t::LIT_STR(..) | t::LIT_STR_RAW(..) => "string",
// number literals
t::LIT_INT(..) | t::LIT_UINT(..) | t::LIT_INT_UNSUFFIXED(..) |
t::LIT_FLOAT(..) | t::LIT_FLOAT_UNSUFFIXED(..) => "number",
t::LIT_INTEGER(..) | t::LIT_FLOAT(..) => "number",
// keywords are also included in the identifier set
t::IDENT(ident, _is_mod_sep) => {

21
src/libsyntax/abi.rs
View file

@ -60,9 +60,12 @@ pub struct AbiData {
}
pub enum AbiArchitecture {
RustArch, // Not a real ABI (e.g., intrinsic)
AllArch, // An ABI that specifies cross-platform defaults (e.g., "C")
Archs(u32) // Multiple architectures (bitset)
/// Not a real ABI (e.g., intrinsic)
RustArch,
/// An ABI that specifies cross-platform defaults (e.g., "C")
AllArch,
/// Multiple architectures (bitset)
Archs(u32)
}
static AbiDatas: &'static [AbiData] = &[
@ -84,21 +87,13 @@ static AbiDatas: &'static [AbiData] = &[
AbiData {abi: RustIntrinsic, name: "rust-intrinsic", abi_arch: RustArch},
];
/// Iterates through each of the defined ABIs.
fn each_abi(op: |abi: Abi| -> bool) -> bool {
/*!
*
* Iterates through each of the defined ABIs.
*/
AbiDatas.iter().advance(|abi_data| op(abi_data.abi))
}
/// Returns the ABI with the given name (if any).
pub fn lookup(name: &str) -> Option<Abi> {
/*!
*
* Returns the ABI with the given name (if any).
*/
let mut res = None;
each_abi(|abi| {

431
src/libsyntax/ast.rs
View file

@ -24,7 +24,8 @@ use std::rc::Rc;
use std::gc::{Gc, GC};
use serialize::{Encodable, Decodable, Encoder, Decoder};
/// A pointer abstraction. FIXME(eddyb) #10676 use Rc<T> in the future.
/// A pointer abstraction.
// FIXME(eddyb) #10676 use Rc<T> in the future.
pub type P<T> = Gc<T>;
#[allow(non_snake_case_functions)]
@ -36,11 +37,11 @@ pub fn P<T: 'static>(value: T) -> P<T> {
// FIXME #6993: in librustc, uses of "ident" should be replaced
// by just "Name".
// an identifier contains a Name (index into the interner
// table) and a SyntaxContext to track renaming and
// macro expansion per Flatt et al., "Macros
// That Work Together"
#[deriving(Clone, Hash, PartialOrd, Eq, Ord, Show)]
/// An identifier contains a Name (index into the interner
/// table) and a SyntaxContext to track renaming and
/// macro expansion per Flatt et al., "Macros
/// That Work Together"
#[deriving(Clone, Hash, PartialOrd, Eq, Ord)]
pub struct Ident {
pub name: Name,
pub ctxt: SyntaxContext
@ -49,6 +50,16 @@ pub struct Ident {
impl Ident {
/// Construct an identifier with the given name and an empty context:
pub fn new(name: Name) -> Ident { Ident {name: name, ctxt: EMPTY_CTXT}}
pub fn as_str<'a>(&'a self) -> &'a str {
self.name.as_str()
}
}
impl Show for Ident {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(f, "\"{}\"#{}", token::get_ident(*self).get(), self.ctxt)
}
}
impl PartialEq for Ident {
@ -95,7 +106,26 @@ pub static ILLEGAL_CTXT : SyntaxContext = 1;
/// A name is a part of an identifier, representing a string or gensym. It's
/// the result of interning.
pub type Name = u32;
#[deriving(Eq, Ord, PartialEq, PartialOrd, Hash, Encodable, Decodable, Clone, Show)]
pub struct Name(pub u32);
impl Name {
pub fn as_str<'a>(&'a self) -> &'a str {
unsafe {
// FIXME #12938: can't use copy_lifetime since &str isn't a &T
::std::mem::transmute(token::get_name(*self).get())
}
}
pub fn uint(&self) -> uint {
let Name(nm) = *self;
nm as uint
}
pub fn ident(&self) -> Ident {
Ident { name: *self, ctxt: 0 }
}
}
/// A mark represents a unique id associated with a macro expansion
pub type Mrk = u32;
@ -122,10 +152,9 @@ pub struct Lifetime {
pub name: Name
}
// a "Path" is essentially Rust's notion of a name;
// for instance: std::cmp::PartialEq . It's represented
// as a sequence of identifiers, along with a bunch
// of supporting information.
/// A "Path" is essentially Rust's notion of a name; for instance:
/// std::cmp::PartialEq . It's represented as a sequence of identifiers,
/// along with a bunch of supporting information.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct Path {
pub span: Span,
@ -163,15 +192,15 @@ pub struct DefId {
pub static LOCAL_CRATE: CrateNum = 0;
pub static CRATE_NODE_ID: NodeId = 0;
// When parsing and doing expansions, we initially give all AST nodes this AST
// node value. Then later, in the renumber pass, we renumber them to have
// small, positive ids.
/// When parsing and doing expansions, we initially give all AST nodes this AST
/// node value. Then later, in the renumber pass, we renumber them to have
/// small, positive ids.
pub static DUMMY_NODE_ID: NodeId = -1;
// The AST represents all type param bounds as types.
// typeck::collect::compute_bounds matches these against
// the "special" built-in traits (see middle::lang_items) and
// detects Copy, Send and Share.
/// The AST represents all type param bounds as types.
/// typeck::collect::compute_bounds matches these against
/// the "special" built-in traits (see middle::lang_items) and
/// detects Copy, Send and Share.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum TyParamBound {
TraitTyParamBound(TraitRef),
@ -210,9 +239,9 @@ impl Generics {
}
}
// The set of MetaItems that define the compilation environment of the crate,
// used to drive conditional compilation
pub type CrateConfig = Vec<Gc<MetaItem>>;
/// The set of MetaItems that define the compilation environment of the crate,
/// used to drive conditional compilation
pub type CrateConfig = Vec<Gc<MetaItem>> ;
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct Crate {
@ -289,13 +318,13 @@ pub enum BindingMode {
pub enum Pat_ {
PatWild,
PatWildMulti,
// A PatIdent may either be a new bound variable,
// or a nullary enum (in which case the third field
// is None).
// In the nullary enum case, the parser can't determine
// which it is. The resolver determines this, and
// records this pattern's NodeId in an auxiliary
// set (of "PatIdents that refer to nullary enums")
/// A PatIdent may either be a new bound variable,
/// or a nullary enum (in which case the third field
/// is None).
/// In the nullary enum case, the parser can't determine
/// which it is. The resolver determines this, and
/// records this pattern's NodeId in an auxiliary
/// set (of "PatIdents that refer to nullary enums")
PatIdent(BindingMode, SpannedIdent, Option<Gc<Pat>>),
PatEnum(Path, Option<Vec<Gc<Pat>>>), /* "none" means a * pattern where
* we don't bind the fields to names */
@ -305,8 +334,8 @@ pub enum Pat_ {
PatRegion(Gc<Pat>), // reference pattern
PatLit(Gc<Expr>),
PatRange(Gc<Expr>, Gc<Expr>),
// [a, b, ..i, y, z] is represented as
// PatVec(~[a, b], Some(i), ~[y, z])
/// [a, b, ..i, y, z] is represented as:
/// PatVec(~[a, b], Some(i), ~[y, z])
PatVec(Vec<Gc<Pat>>, Option<Gc<Pat>>, Vec<Gc<Pat>>),
PatMac(Mac),
}
@ -319,9 +348,12 @@ pub enum Mutability {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum ExprVstore {
ExprVstoreUniq, // ~[1,2,3,4]
ExprVstoreSlice, // &[1,2,3,4]
ExprVstoreMutSlice, // &mut [1,2,3,4]
/// ~[1, 2, 3, 4]
ExprVstoreUniq,
/// &[1, 2, 3, 4]
ExprVstoreSlice,
/// &mut [1, 2, 3, 4]
ExprVstoreMutSlice,
}
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
@ -359,16 +391,16 @@ pub type Stmt = Spanned<Stmt_>;
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum Stmt_ {
// could be an item or a local (let) binding:
/// Could be an item or a local (let) binding:
StmtDecl(Gc<Decl>, NodeId),
// expr without trailing semi-colon (must have unit type):
/// Expr without trailing semi-colon (must have unit type):
StmtExpr(Gc<Expr>, NodeId),
// expr with trailing semi-colon (may have any type):
/// Expr with trailing semi-colon (may have any type):
StmtSemi(Gc<Expr>, NodeId),
// bool: is there a trailing sem-colon?
/// bool: is there a trailing sem-colon?
StmtMac(Mac, bool),
}
@ -397,9 +429,9 @@ pub type Decl = Spanned<Decl_>;
#[deriving(PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum Decl_ {
// a local (let) binding:
/// A local (let) binding:
DeclLocal(Gc<Local>),
// an item binding:
/// An item binding:
DeclItem(Gc<Item>),
}
@ -443,7 +475,7 @@ pub struct Expr {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum Expr_ {
ExprVstore(Gc<Expr>, ExprVstore),
// First expr is the place; second expr is the value.
/// First expr is the place; second expr is the value.
ExprBox(Gc<Expr>, Gc<Expr>),
ExprVec(Vec<Gc<Expr>>),
ExprCall(Gc<Expr>, Vec<Gc<Expr>>),
@ -483,130 +515,127 @@ pub enum Expr_ {
ExprMac(Mac),
// A struct literal expression.
/// A struct literal expression.
ExprStruct(Path, Vec<Field> , Option<Gc<Expr>> /* base */),
// A vector literal constructed from one repeated element.
/// A vector literal constructed from one repeated element.
ExprRepeat(Gc<Expr> /* element */, Gc<Expr> /* count */),
// No-op: used solely so we can pretty-print faithfully
/// No-op: used solely so we can pretty-print faithfully
ExprParen(Gc<Expr>)
}
// When the main rust parser encounters a syntax-extension invocation, it
// parses the arguments to the invocation as a token-tree. This is a very
// loose structure, such that all sorts of different AST-fragments can
// be passed to syntax extensions using a uniform type.
//
// If the syntax extension is an MBE macro, it will attempt to match its
// LHS "matchers" against the provided token tree, and if it finds a
// match, will transcribe the RHS token tree, splicing in any captured
// macro_parser::matched_nonterminals into the TTNonterminals it finds.
//
// The RHS of an MBE macro is the only place a TTNonterminal or TTSeq
// makes any real sense. You could write them elsewhere but nothing
// else knows what to do with them, so you'll probably get a syntax
// error.
//
/// When the main rust parser encounters a syntax-extension invocation, it
/// parses the arguments to the invocation as a token-tree. This is a very
/// loose structure, such that all sorts of different AST-fragments can
/// be passed to syntax extensions using a uniform type.
///
/// If the syntax extension is an MBE macro, it will attempt to match its
/// LHS "matchers" against the provided token tree, and if it finds a
/// match, will transcribe the RHS token tree, splicing in any captured
/// macro_parser::matched_nonterminals into the TTNonterminals it finds.
///
/// The RHS of an MBE macro is the only place a TTNonterminal or TTSeq
/// makes any real sense. You could write them elsewhere but nothing
/// else knows what to do with them, so you'll probably get a syntax
/// error.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
#[doc="For macro invocations; parsing is delegated to the macro"]
pub enum TokenTree {
// a single token
/// A single token
TTTok(Span, ::parse::token::Token),
// a delimited sequence (the delimiters appear as the first
// and last elements of the vector)
/// A delimited sequence (the delimiters appear as the first
/// and last elements of the vector)
// FIXME(eddyb) #6308 Use Rc<[TokenTree]> after DST.
TTDelim(Rc<Vec<TokenTree>>),
// These only make sense for right-hand-sides of MBE macros:
// a kleene-style repetition sequence with a span, a TTForest,
// an optional separator, and a boolean where true indicates
// zero or more (..), and false indicates one or more (+).
/// A kleene-style repetition sequence with a span, a TTForest,
/// an optional separator, and a boolean where true indicates
/// zero or more (..), and false indicates one or more (+).
// FIXME(eddyb) #6308 Use Rc<[TokenTree]> after DST.
TTSeq(Span, Rc<Vec<TokenTree>>, Option<::parse::token::Token>, bool),
// a syntactic variable that will be filled in by macro expansion.
/// A syntactic variable that will be filled in by macro expansion.
TTNonterminal(Span, Ident)
}
//
// Matchers are nodes defined-by and recognized-by the main rust parser and
// language, but they're only ever found inside syntax-extension invocations;
// indeed, the only thing that ever _activates_ the rules in the rust parser
// for parsing a matcher is a matcher looking for the 'matchers' nonterminal
// itself. Matchers represent a small sub-language for pattern-matching
// token-trees, and are thus primarily used by the macro-defining extension
// itself.
//
// MatchTok
// --------
//
// A matcher that matches a single token, denoted by the token itself. So
// long as there's no $ involved.
//
//
// MatchSeq
// --------
//
// A matcher that matches a sequence of sub-matchers, denoted various
// possible ways:
//
// $(M)* zero or more Ms
// $(M)+ one or more Ms
// $(M),+ one or more comma-separated Ms
// $(A B C);* zero or more semi-separated 'A B C' seqs
//
//
// MatchNonterminal
// -----------------
//
// A matcher that matches one of a few interesting named rust
// nonterminals, such as types, expressions, items, or raw token-trees. A
// black-box matcher on expr, for example, binds an expr to a given ident,
// and that ident can re-occur as an interpolation in the RHS of a
// macro-by-example rule. For example:
//
// $foo:expr => 1 + $foo // interpolate an expr
// $foo:tt => $foo // interpolate a token-tree
// $foo:tt => bar! $foo // only other valid interpolation
// // is in arg position for another
// // macro
//
// As a final, horrifying aside, note that macro-by-example's input is
// also matched by one of these matchers. Holy self-referential! It is matched
// by a MatchSeq, specifically this one:
//
// $( $lhs:matchers => $rhs:tt );+
//
// If you understand that, you have closed to loop and understand the whole
// macro system. Congratulations.
//
/// Matchers are nodes defined-by and recognized-by the main rust parser and
/// language, but they're only ever found inside syntax-extension invocations;
/// indeed, the only thing that ever _activates_ the rules in the rust parser
/// for parsing a matcher is a matcher looking for the 'matchers' nonterminal
/// itself. Matchers represent a small sub-language for pattern-matching
/// token-trees, and are thus primarily used by the macro-defining extension
/// itself.
///
/// MatchTok
/// --------
///
/// A matcher that matches a single token, denoted by the token itself. So
/// long as there's no $ involved.
///
///
/// MatchSeq
/// --------
///
/// A matcher that matches a sequence of sub-matchers, denoted various
/// possible ways:
///
/// $(M)* zero or more Ms
/// $(M)+ one or more Ms
/// $(M),+ one or more comma-separated Ms
/// $(A B C);* zero or more semi-separated 'A B C' seqs
///
///
/// MatchNonterminal
/// -----------------
///
/// A matcher that matches one of a few interesting named rust
/// nonterminals, such as types, expressions, items, or raw token-trees. A
/// black-box matcher on expr, for example, binds an expr to a given ident,
/// and that ident can re-occur as an interpolation in the RHS of a
/// macro-by-example rule. For example:
///
/// $foo:expr => 1 + $foo // interpolate an expr
/// $foo:tt => $foo // interpolate a token-tree
/// $foo:tt => bar! $foo // only other valid interpolation
/// // is in arg position for another
/// // macro
///
/// As a final, horrifying aside, note that macro-by-example's input is
/// also matched by one of these matchers. Holy self-referential! It is matched
/// by a MatchSeq, specifically this one:
///
/// $( $lhs:matchers => $rhs:tt );+
///
/// If you understand that, you have closed the loop and understand the whole
/// macro system. Congratulations.
pub type Matcher = Spanned<Matcher_>;
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum Matcher_ {
// match one token
/// Match one token
MatchTok(::parse::token::Token),
// match repetitions of a sequence: body, separator, zero ok?,
// lo, hi position-in-match-array used:
/// Match repetitions of a sequence: body, separator, zero ok?,
/// lo, hi position-in-match-array used:
MatchSeq(Vec<Matcher> , Option<::parse::token::Token>, bool, uint, uint),
// parse a Rust NT: name to bind, name of NT, position in match array:
/// Parse a Rust NT: name to bind, name of NT, position in match array:
MatchNonterminal(Ident, Ident, uint)
}
pub type Mac = Spanned<Mac_>;
// represents a macro invocation. The Path indicates which macro
// is being invoked, and the vector of token-trees contains the source
// of the macro invocation.
// There's only one flavor, now, so this could presumably be simplified.
/// Represents a macro invocation. The Path indicates which macro
/// is being invoked, and the vector of token-trees contains the source
/// of the macro invocation.
/// There's only one flavor, now, so this could presumably be simplified.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum Mac_ {
MacInvocTT(Path, Vec<TokenTree> , SyntaxContext), // new macro-invocation
}
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash, Show)]
pub enum StrStyle {
CookedStr,
RawStr(uint)
@ -614,7 +643,7 @@ pub enum StrStyle {
pub type Lit = Spanned<Lit_>;
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash, Show)]
pub enum Lit_ {
LitStr(InternedString, StrStyle),
LitBinary(Rc<Vec<u8> >),
@ -659,11 +688,10 @@ pub struct TypeMethod {
pub vis: Visibility,
}
/// Represents a method declaration in a trait declaration, possibly
/// including a default implementation
// A trait method is either required (meaning it doesn't have an
// implementation, just a signature) or provided (meaning it has a default
// implementation).
/// Represents a method declaration in a trait declaration, possibly including
/// a default implementation A trait method is either required (meaning it
/// doesn't have an implementation, just a signature) or provided (meaning it
/// has a default implementation).
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum TraitMethod {
Required(TypeMethod),
@ -685,6 +713,16 @@ impl fmt::Show for IntTy {
}
}
impl IntTy {
pub fn suffix_len(&self) -> uint {
match *self {
TyI => 1,
TyI8 => 2,
TyI16 | TyI32 | TyI64 => 3,
}
}
}
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum UintTy {
TyU,
@ -694,6 +732,16 @@ pub enum UintTy {
TyU64,
}
impl UintTy {
pub fn suffix_len(&self) -> uint {
match *self {
TyU => 1,
TyU8 => 2,
TyU16 | TyU32 | TyU64 => 3,
}
}
}
impl fmt::Show for UintTy {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(f, "{}", ast_util::uint_ty_to_string(*self, None))
@ -712,6 +760,14 @@ impl fmt::Show for FloatTy {
}
}
impl FloatTy {
pub fn suffix_len(&self) -> uint {
match *self {
TyF32 | TyF64 => 3, // add F128 handling here
}
}
}
// NB PartialEq method appears below.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct Ty {
@ -720,7 +776,7 @@ pub struct Ty {
pub span: Span,
}
// Not represented directly in the AST, referred to by name through a ty_path.
/// Not represented directly in the AST, referred to by name through a ty_path.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum PrimTy {
TyInt(IntTy),
@ -753,10 +809,10 @@ pub struct ClosureTy {
pub fn_style: FnStyle,
pub onceness: Onceness,
pub decl: P<FnDecl>,
// Optional optvec distinguishes between "fn()" and "fn:()" so we can
// implement issue #7264. None means "fn()", which means infer a default
// bound based on pointer sigil during typeck. Some(Empty) means "fn:()",
// which means use no bounds (e.g., not even Owned on a ~fn()).
/// Optional optvec distinguishes between "fn()" and "fn:()" so we can
/// implement issue #7264. None means "fn()", which means infer a default
/// bound based on pointer sigil during typeck. Some(Empty) means "fn:()",
/// which means use no bounds (e.g., not even Owned on a ~fn()).
pub bounds: Option<OwnedSlice<TyParamBound>>,
}
@ -789,11 +845,11 @@ pub enum Ty_ {
TyUnboxedFn(Gc<UnboxedFnTy>),
TyTup(Vec<P<Ty>> ),
TyPath(Path, Option<OwnedSlice<TyParamBound>>, NodeId), // for #7264; see above
// No-op; kept solely so that we can pretty-print faithfully
/// No-op; kept solely so that we can pretty-print faithfully
TyParen(P<Ty>),
TyTypeof(Gc<Expr>),
// TyInfer means the type should be inferred instead of it having been
// specified. This can appear anywhere in a type.
/// TyInfer means the type should be inferred instead of it having been
/// specified. This can appear anywhere in a type.
TyInfer,
}
@ -854,8 +910,10 @@ pub struct FnDecl {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum FnStyle {
UnsafeFn, // declared with "unsafe fn"
NormalFn, // declared with "fn"
/// Declared with "unsafe fn"
UnsafeFn,
/// Declared with "fn"
NormalFn,
}
impl fmt::Show for FnStyle {
@ -869,18 +927,24 @@ impl fmt::Show for FnStyle {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum RetStyle {
NoReturn, // functions with return type _|_ that always
// raise an error or exit (i.e. never return to the caller)
Return, // everything else
/// Functions with return type ! that always
/// raise an error or exit (i.e. never return to the caller)
NoReturn,
/// Everything else
Return,
}
/// Represents the kind of 'self' associated with a method
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum ExplicitSelf_ {
SelfStatic, // no self
SelfValue(Ident), // `self`
SelfRegion(Option<Lifetime>, Mutability, Ident), // `&'lt self`, `&'lt mut self`
SelfUniq(Ident), // `~self`
/// No self
SelfStatic,
/// `self
SelfValue(Ident),
/// `&'lt self`, `&'lt mut self`
SelfRegion(Option<Lifetime>, Mutability, Ident),
/// `~self`
SelfUniq(Ident)
}
pub type ExplicitSelf = Spanned<ExplicitSelf_>;
@ -959,17 +1023,17 @@ pub type ViewPath = Spanned<ViewPath_>;
#[deriving(PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum ViewPath_ {
// quux = foo::bar::baz
//
// or just
//
// foo::bar::baz (with 'baz =' implicitly on the left)
/// `quux = foo::bar::baz`
///
/// or just
///
/// `foo::bar::baz ` (with 'baz =' implicitly on the left)
ViewPathSimple(Ident, Path, NodeId),
// foo::bar::*
/// `foo::bar::*`
ViewPathGlob(Path, NodeId),
// foo::bar::{a,b,c}
/// `foo::bar::{a,b,c}`
ViewPathList(Path, Vec<PathListIdent> , NodeId)
}
@ -983,20 +1047,20 @@ pub struct ViewItem {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum ViewItem_ {
// ident: name used to refer to this crate in the code
// optional (InternedString,StrStyle): if present, this is a location
// (containing arbitrary characters) from which to fetch the crate sources
// For example, extern crate whatever = "github.com/rust-lang/rust"
/// Ident: name used to refer to this crate in the code
/// optional (InternedString,StrStyle): if present, this is a location
/// (containing arbitrary characters) from which to fetch the crate sources
/// For example, extern crate whatever = "github.com/rust-lang/rust"
ViewItemExternCrate(Ident, Option<(InternedString,StrStyle)>, NodeId),
ViewItemUse(Gc<ViewPath>),
}
// Meta-data associated with an item
/// Meta-data associated with an item
pub type Attribute = Spanned<Attribute_>;
// Distinguishes between Attributes that decorate items and Attributes that
// are contained as statements within items. These two cases need to be
// distinguished for pretty-printing.
/// Distinguishes between Attributes that decorate items and Attributes that
/// are contained as statements within items. These two cases need to be
/// distinguished for pretty-printing.
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum AttrStyle {
AttrOuter,
@ -1006,7 +1070,7 @@ pub enum AttrStyle {
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct AttrId(pub uint);
// doc-comments are promoted to attributes that have is_sugared_doc = true
/// Doc-comments are promoted to attributes that have is_sugared_doc = true
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct Attribute_ {
pub id: AttrId,
@ -1015,13 +1079,12 @@ pub struct Attribute_ {
pub is_sugared_doc: bool,
}
/*
TraitRef's appear in impls.
resolve maps each TraitRef's ref_id to its defining trait; that's all
that the ref_id is for. The impl_id maps to the "self type" of this impl.
If this impl is an ItemImpl, the impl_id is redundant (it could be the
same as the impl's node id).
*/
/// TraitRef's appear in impls.
/// resolve maps each TraitRef's ref_id to its defining trait; that's all
/// that the ref_id is for. The impl_id maps to the "self type" of this impl.
/// If this impl is an ItemImpl, the impl_id is redundant (it could be the
/// same as the impl's node id).
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct TraitRef {
pub path: Path,
@ -1065,7 +1128,8 @@ pub type StructField = Spanned<StructField_>;
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum StructFieldKind {
NamedField(Ident, Visibility),
UnnamedField(Visibility), // element of a tuple-like struct
/// Element of a tuple-like struct
UnnamedField(Visibility),
}
impl StructFieldKind {
@ -1079,12 +1143,15 @@ impl StructFieldKind {
#[deriving(PartialEq, Eq, Encodable, Decodable, Hash)]
pub struct StructDef {
pub fields: Vec<StructField>, /* fields, not including ctor */
/* ID of the constructor. This is only used for tuple- or enum-like
* structs. */
/// Fields, not including ctor
pub fields: Vec<StructField>,
/// ID of the constructor. This is only used for tuple- or enum-like
/// structs.
pub ctor_id: Option<NodeId>,
pub super_struct: Option<P<Ty>>, // Super struct, if specified.
pub is_virtual: bool, // True iff the struct may be inherited from.
/// Super struct, if specified.
pub super_struct: Option<P<Ty>>,
/// True iff the struct may be inherited from.
pub is_virtual: bool,
}
/*
@ -1120,7 +1187,7 @@ pub enum Item_ {
Option<TraitRef>, // (optional) trait this impl implements
P<Ty>, // self
Vec<Gc<Method>>),
// a macro invocation (which includes macro definition)
/// A macro invocation (which includes macro definition)
ItemMac(Mac),
}
@ -1140,9 +1207,9 @@ pub enum ForeignItem_ {
ForeignItemStatic(P<Ty>, /* is_mutbl */ bool),
}
// The data we save and restore about an inlined item or method. This is not
// part of the AST that we parse from a file, but it becomes part of the tree
// that we trans.
/// The data we save and restore about an inlined item or method. This is not
/// part of the AST that we parse from a file, but it becomes part of the tree
/// that we trans.
#[deriving(PartialEq, Eq, Encodable, Decodable, Hash)]
pub enum InlinedItem {
IIItem(Gc<Item>),

22
src/libsyntax/ast_map.rs
View file

@ -112,13 +112,13 @@ pub enum Node {
NodeLifetime(Gc<Lifetime>),
}
// The odd layout is to bring down the total size.
/// The odd layout is to bring down the total size.
#[deriving(Clone)]
enum MapEntry {
// Placeholder for holes in the map.
/// Placeholder for holes in the map.
NotPresent,
// All the node types, with a parent ID.
/// All the node types, with a parent ID.
EntryItem(NodeId, Gc<Item>),
EntryForeignItem(NodeId, Gc<ForeignItem>),
EntryTraitMethod(NodeId, Gc<TraitMethod>),
@ -133,14 +133,14 @@ enum MapEntry {
EntryStructCtor(NodeId, Gc<StructDef>),
EntryLifetime(NodeId, Gc<Lifetime>),
// Roots for node trees.
/// Roots for node trees.
RootCrate,
RootInlinedParent(P<InlinedParent>)
}
struct InlinedParent {
path: Vec<PathElem> ,
// Required by NodeTraitMethod and NodeMethod.
/// Required by NodeTraitMethod and NodeMethod.
def_id: DefId
}
@ -243,7 +243,7 @@ impl Map {
ItemForeignMod(ref nm) => Some(nm.abi),
_ => None
},
// Wrong but OK, because the only inlined foreign items are intrinsics.
/// Wrong but OK, because the only inlined foreign items are intrinsics.
Some(RootInlinedParent(_)) => Some(abi::RustIntrinsic),
_ => None
};
@ -432,8 +432,8 @@ pub trait FoldOps {
pub struct Ctx<'a, F> {
map: &'a Map,
// The node in which we are currently mapping (an item or a method).
// When equal to DUMMY_NODE_ID, the next mapped node becomes the parent.
/// The node in which we are currently mapping (an item or a method).
/// When equal to DUMMY_NODE_ID, the next mapped node becomes the parent.
parent: NodeId,
fold_ops: F
}
@ -618,9 +618,9 @@ pub fn map_crate<F: FoldOps>(krate: Crate, fold_ops: F) -> (Crate, Map) {
(krate, map)
}
// Used for items loaded from external crate that are being inlined into this
// crate. The `path` should be the path to the item but should not include
// the item itself.
/// Used for items loaded from external crate that are being inlined into this
/// crate. The `path` should be the path to the item but should not include
/// the item itself.
pub fn map_decoded_item<F: FoldOps>(map: &Map,
path: Vec<PathElem> ,
fold_ops: F,

22
src/libsyntax/ast_util.rs
View file

@ -101,8 +101,8 @@ pub fn is_path(e: Gc<Expr>) -> bool {
return match e.node { ExprPath(_) => true, _ => false };
}
// Get a string representation of a signed int type, with its value.
// We want to avoid "45int" and "-3int" in favor of "45" and "-3"
/// Get a string representation of a signed int type, with its value.
/// We want to avoid "45int" and "-3int" in favor of "45" and "-3"
pub fn int_ty_to_string(t: IntTy, val: Option<i64>) -> String {
let s = match t {
TyI if val.is_some() => "i",
@ -131,8 +131,8 @@ pub fn int_ty_max(t: IntTy) -> u64 {
}
}
// Get a string representation of an unsigned int type, with its value.
// We want to avoid "42uint" in favor of "42u"
/// Get a string representation of an unsigned int type, with its value.
/// We want to avoid "42uint" in favor of "42u"
pub fn uint_ty_to_string(t: UintTy, val: Option<u64>) -> String {
let s = match t {
TyU if val.is_some() => "u",
@ -249,8 +249,8 @@ pub fn public_methods(ms: Vec<Gc<Method>> ) -> Vec<Gc<Method>> {
}).collect()
}
// extract a TypeMethod from a TraitMethod. if the TraitMethod is
// a default, pull out the useful fields to make a TypeMethod
/// extract a TypeMethod from a TraitMethod. if the TraitMethod is
/// a default, pull out the useful fields to make a TypeMethod
pub fn trait_method_to_ty_method(method: &TraitMethod) -> TypeMethod {
match *method {
Required(ref m) => (*m).clone(),
@ -705,7 +705,7 @@ pub fn segments_name_eq(a : &[ast::PathSegment], b : &[ast::PathSegment]) -> boo
}
}
// Returns true if this literal is a string and false otherwise.
/// Returns true if this literal is a string and false otherwise.
pub fn lit_is_str(lit: Gc<Lit>) -> bool {
match lit.node {
LitStr(..) => true,
@ -754,14 +754,14 @@ mod test {
#[test] fn idents_name_eq_test() {
assert!(segments_name_eq(
[Ident{name:3,ctxt:4}, Ident{name:78,ctxt:82}]
[Ident{name:Name(3),ctxt:4}, Ident{name:Name(78),ctxt:82}]
.iter().map(ident_to_segment).collect::<Vec<PathSegment>>().as_slice(),
[Ident{name:3,ctxt:104}, Ident{name:78,ctxt:182}]
[Ident{name:Name(3),ctxt:104}, Ident{name:Name(78),ctxt:182}]
.iter().map(ident_to_segment).collect::<Vec<PathSegment>>().as_slice()));
assert!(!segments_name_eq(
[Ident{name:3,ctxt:4}, Ident{name:78,ctxt:82}]
[Ident{name:Name(3),ctxt:4}, Ident{name:Name(78),ctxt:82}]
.iter().map(ident_to_segment).collect::<Vec<PathSegment>>().as_slice(),
[Ident{name:3,ctxt:104}, Ident{name:77,ctxt:182}]
[Ident{name:Name(3),ctxt:104}, Ident{name:Name(77),ctxt:182}]
.iter().map(ident_to_segment).collect::<Vec<PathSegment>>().as_slice()));
}
}

28
src/libsyntax/attr.rs
View file

@ -46,10 +46,8 @@ pub trait AttrMetaMethods {
/// #[foo="bar"] and #[foo(bar)]
fn name(&self) -> InternedString;
/**
* Gets the string value if self is a MetaNameValue variant
* containing a string, otherwise None.
*/
/// Gets the string value if self is a MetaNameValue variant
/// containing a string, otherwise None.
fn value_str(&self) -> Option<InternedString>;
/// Gets a list of inner meta items from a list MetaItem type.
fn meta_item_list<'a>(&'a self) -> Option<&'a [Gc<MetaItem>]>;
@ -420,18 +418,16 @@ pub fn require_unique_names(diagnostic: &SpanHandler, metas: &[Gc<MetaItem>]) {
}
/**
* Fold this over attributes to parse #[repr(...)] forms.
*
* Valid repr contents: any of the primitive integral type names (see
* `int_type_of_word`, below) to specify the discriminant type; and `C`, to use
* the same discriminant size that the corresponding C enum would. These are
* not allowed on univariant or zero-variant enums, which have no discriminant.
*
* If a discriminant type is so specified, then the discriminant will be
* present (before fields, if any) with that type; reprensentation
* optimizations which would remove it will not be done.
*/
/// Fold this over attributes to parse #[repr(...)] forms.
///
/// Valid repr contents: any of the primitive integral type names (see
/// `int_type_of_word`, below) to specify the discriminant type; and `C`, to use
/// the same discriminant size that the corresponding C enum would. These are
/// not allowed on univariant or zero-variant enums, which have no discriminant.
///
/// If a discriminant type is so specified, then the discriminant will be
/// present (before fields, if any) with that type; reprensentation
/// optimizations which would remove it will not be done.
pub fn find_repr_attr(diagnostic: &SpanHandler, attr: &Attribute, acc: ReprAttr)
-> ReprAttr {
let mut acc = acc;

33
src/libsyntax/codemap.rs
View file

@ -96,7 +96,7 @@ pub struct Span {
pub static DUMMY_SP: Span = Span { lo: BytePos(0), hi: BytePos(0), expn_info: None };
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash)]
#[deriving(Clone, PartialEq, Eq, Encodable, Decodable, Hash, Show)]
pub struct Spanned<T> {
pub node: T,
pub span: Span,
@ -252,15 +252,15 @@ pub struct FileMap {
}
impl FileMap {
// EFFECT: register a start-of-line offset in the
// table of line-beginnings.
// UNCHECKED INVARIANT: these offsets must be added in the right
// order and must be in the right places; there is shared knowledge
// about what ends a line between this file and parse.rs
// WARNING: pos param here is the offset relative to start of CodeMap,
// and CodeMap will append a newline when adding a filemap without a newline at the end,
// so the safe way to call this is with value calculated as
// filemap.start_pos + newline_offset_relative_to_the_start_of_filemap.
/// EFFECT: register a start-of-line offset in the
/// table of line-beginnings.
/// UNCHECKED INVARIANT: these offsets must be added in the right
/// order and must be in the right places; there is shared knowledge
/// about what ends a line between this file and parse.rs
/// WARNING: pos param here is the offset relative to start of CodeMap,
/// and CodeMap will append a newline when adding a filemap without a newline at the end,
/// so the safe way to call this is with value calculated as
/// filemap.start_pos + newline_offset_relative_to_the_start_of_filemap.
pub fn next_line(&self, pos: BytePos) {
// the new charpos must be > the last one (or it's the first one).
let mut lines = self.lines.borrow_mut();;
@ -269,7 +269,7 @@ impl FileMap {
lines.push(pos);
}
// get a line from the list of pre-computed line-beginnings
/// get a line from the list of pre-computed line-beginnings
pub fn get_line(&self, line: int) -> String {
let mut lines = self.lines.borrow_mut();
let begin: BytePos = *lines.get(line as uint) - self.start_pos;
@ -428,9 +428,8 @@ impl CodeMap {
FileMapAndBytePos {fm: fm, pos: offset}
}
// Converts an absolute BytePos to a CharPos relative to the filemap and above.
/// Converts an absolute BytePos to a CharPos relative to the filemap and above.
pub fn bytepos_to_file_charpos(&self, bpos: BytePos) -> CharPos {
debug!("codemap: converting {:?} to char pos", bpos);
let idx = self.lookup_filemap_idx(bpos);
let files = self.files.borrow();
let map = files.get(idx);
@ -439,7 +438,7 @@ impl CodeMap {
let mut total_extra_bytes = 0;
for mbc in map.multibyte_chars.borrow().iter() {
debug!("codemap: {:?}-byte char at {:?}", mbc.bytes, mbc.pos);
debug!("{}-byte char at {}", mbc.bytes, mbc.pos);
if mbc.pos < bpos {
// every character is at least one byte, so we only
// count the actual extra bytes.
@ -514,11 +513,11 @@ impl CodeMap {
let chpos = self.bytepos_to_file_charpos(pos);
let linebpos = *f.lines.borrow().get(a);
let linechpos = self.bytepos_to_file_charpos(linebpos);
debug!("codemap: byte pos {:?} is on the line at byte pos {:?}",
debug!("byte pos {} is on the line at byte pos {}",
pos, linebpos);
debug!("codemap: char pos {:?} is on the line at char pos {:?}",
debug!("char pos {} is on the line at char pos {}",
chpos, linechpos);
debug!("codemap: byte is on line: {:?}", line);
debug!("byte is on line: {}", line);
assert!(chpos >= linechpos);
Loc {
file: f,

26
src/libsyntax/diagnostic.rs
View file

@ -21,7 +21,7 @@ use std::string::String;
use term::WriterWrapper;
use term;
// maximum number of lines we will print for each error; arbitrary.
/// maximum number of lines we will print for each error; arbitrary.
static MAX_LINES: uint = 6u;
#[deriving(Clone)]
@ -73,9 +73,9 @@ pub struct FatalError;
/// or `.span_bug` rather than a failed assertion, etc.
pub struct ExplicitBug;
// a span-handler is like a handler but also
// accepts span information for source-location
// reporting.
/// A span-handler is like a handler but also
/// accepts span information for source-location
/// reporting.
pub struct SpanHandler {
pub handler: Handler,
pub cm: codemap::CodeMap,
@ -114,9 +114,9 @@ impl SpanHandler {
}
}
// a handler deals with errors; certain errors
// (fatal, bug, unimpl) may cause immediate exit,
// others log errors for later reporting.
/// A handler deals with errors; certain errors
/// (fatal, bug, unimpl) may cause immediate exit,
/// others log errors for later reporting.
pub struct Handler {
err_count: Cell<uint>,
emit: RefCell<Box<Emitter + Send>>,
@ -442,12 +442,12 @@ fn highlight_lines(err: &mut EmitterWriter,
Ok(())
}
// Here are the differences between this and the normal `highlight_lines`:
// `custom_highlight_lines` will always put arrow on the last byte of the
// span (instead of the first byte). Also, when the span is too long (more
// than 6 lines), `custom_highlight_lines` will print the first line, then
// dot dot dot, then last line, whereas `highlight_lines` prints the first
// six lines.
/// Here are the differences between this and the normal `highlight_lines`:
/// `custom_highlight_lines` will always put arrow on the last byte of the
/// span (instead of the first byte). Also, when the span is too long (more
/// than 6 lines), `custom_highlight_lines` will print the first line, then
/// dot dot dot, then last line, whereas `highlight_lines` prints the first
/// six lines.
fn custom_highlight_lines(w: &mut EmitterWriter,
cm: &codemap::CodeMap,
sp: Span,

32
src/libsyntax/ext/base.rs
View file

@ -278,9 +278,9 @@ pub enum SyntaxExtension {
pub type NamedSyntaxExtension = (Name, SyntaxExtension);
pub struct BlockInfo {
// should macros escape from this scope?
/// Should macros escape from this scope?
pub macros_escape: bool,
// what are the pending renames?
/// What are the pending renames?
pub pending_renames: mtwt::RenameList,
}
@ -293,8 +293,8 @@ impl BlockInfo {
}
}
// The base map of methods for expanding syntax extension
// AST nodes into full ASTs
/// The base map of methods for expanding syntax extension
/// AST nodes into full ASTs
pub fn syntax_expander_table() -> SyntaxEnv {
// utility function to simplify creating NormalTT syntax extensions
fn builtin_normal_expander(f: MacroExpanderFn) -> SyntaxExtension {
@ -398,9 +398,9 @@ pub fn syntax_expander_table() -> SyntaxEnv {
syntax_expanders
}
// One of these is made during expansion and incrementally updated as we go;
// when a macro expansion occurs, the resulting nodes have the backtrace()
// -> expn_info of their expansion context stored into their span.
/// One of these is made during expansion and incrementally updated as we go;
/// when a macro expansion occurs, the resulting nodes have the backtrace()
/// -> expn_info of their expansion context stored into their span.
pub struct ExtCtxt<'a> {
pub parse_sess: &'a parse::ParseSess,
pub cfg: ast::CrateConfig,
@ -535,6 +535,9 @@ impl<'a> ExtCtxt<'a> {
pub fn ident_of(&self, st: &str) -> ast::Ident {
str_to_ident(st)
}
pub fn name_of(&self, st: &str) -> ast::Name {
token::intern(st)
}
}
/// Extract a string literal from the macro expanded version of `expr`,
@ -579,9 +582,9 @@ pub fn get_single_str_from_tts(cx: &ExtCtxt,
cx.span_err(sp, format!("{} takes 1 argument.", name).as_slice());
} else {
match tts[0] {
ast::TTTok(_, token::LIT_STR(ident))
| ast::TTTok(_, token::LIT_STR_RAW(ident, _)) => {
return Some(token::get_ident(ident).get().to_string())
ast::TTTok(_, token::LIT_STR(ident)) => return Some(parse::str_lit(ident.as_str())),
ast::TTTok(_, token::LIT_STR_RAW(ident, _)) => {
return Some(parse::raw_str_lit(ident.as_str()))
}
_ => {
cx.span_err(sp,
@ -612,11 +615,11 @@ pub fn get_exprs_from_tts(cx: &mut ExtCtxt,
Some(es)
}
// in order to have some notion of scoping for macros,
// we want to implement the notion of a transformation
// environment.
/// In order to have some notion of scoping for macros,
/// we want to implement the notion of a transformation
/// environment.
// This environment maps Names to SyntaxExtensions.
/// This environment maps Names to SyntaxExtensions.
//impl question: how to implement it? Initially, the
// env will contain only macros, so it might be painful
@ -633,7 +636,6 @@ struct MapChainFrame {
map: HashMap<Name, SyntaxExtension>,
}
// Only generic to make it easy to test
pub struct SyntaxEnv {
chain: Vec<MapChainFrame> ,
}

143
src/libsyntax/ext/deriving/encodable.rs
View file

@ -8,79 +8,76 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
/*!
The compiler code necessary to implement the `#[deriving(Encodable)]`
(and `Decodable`, in decodable.rs) extension. The idea here is that
type-defining items may be tagged with `#[deriving(Encodable, Decodable)]`.
For example, a type like:
```ignore
#[deriving(Encodable, Decodable)]
struct Node { id: uint }
```
would generate two implementations like:
```ignore
impl<S:serialize::Encoder> Encodable<S> for Node {
fn encode(&self, s: &S) {
s.emit_struct("Node", 1, || {
s.emit_field("id", 0, || s.emit_uint(self.id))
})
}
}
impl<D:Decoder> Decodable for node_id {
fn decode(d: &D) -> Node {
d.read_struct("Node", 1, || {
Node {
id: d.read_field("x".to_string(), 0, || decode(d))
}
})
}
}
```
Other interesting scenarios are whe the item has type parameters or
references other non-built-in types. A type definition like:
```ignore
#[deriving(Encodable, Decodable)]
struct spanned<T> { node: T, span: Span }
```
would yield functions like:
```ignore
impl<
S: Encoder,
T: Encodable<S>
> spanned<T>: Encodable<S> {
fn encode<S:Encoder>(s: &S) {
s.emit_rec(|| {
s.emit_field("node", 0, || self.node.encode(s));
s.emit_field("span", 1, || self.span.encode(s));
})
}
}
impl<
D: Decoder,
T: Decodable<D>
> spanned<T>: Decodable<D> {
fn decode(d: &D) -> spanned<T> {
d.read_rec(|| {
{
node: d.read_field("node".to_string(), 0, || decode(d)),
span: d.read_field("span".to_string(), 1, || decode(d)),
}
})
}
}
```
*/
//! The compiler code necessary to implement the `#[deriving(Encodable)]`
//! (and `Decodable`, in decodable.rs) extension. The idea here is that
//! type-defining items may be tagged with `#[deriving(Encodable, Decodable)]`.
//!
//! For example, a type like:
//!
//! ```ignore
//! #[deriving(Encodable, Decodable)]
//! struct Node { id: uint }
//! ```
//!
//! would generate two implementations like:
//!
//! ```ignore
//! impl<S:serialize::Encoder> Encodable<S> for Node {
//! fn encode(&self, s: &S) {
//! s.emit_struct("Node", 1, || {
//! s.emit_field("id", 0, || s.emit_uint(self.id))
//! })
//! }
//! }
//!
//! impl<D:Decoder> Decodable for node_id {
//! fn decode(d: &D) -> Node {
//! d.read_struct("Node", 1, || {
//! Node {
//! id: d.read_field("x".to_string(), 0, || decode(d))
//! }
//! })
//! }
//! }
//! ```
//!
//! Other interesting scenarios are whe the item has type parameters or
//! references other non-built-in types. A type definition like:
//!
//! ```ignore
//! #[deriving(Encodable, Decodable)]
//! struct spanned<T> { node: T, span: Span }
//! ```
//!
//! would yield functions like:
//!
//! ```ignore
//! impl<
//! S: Encoder,
//! T: Encodable<S>
//! > spanned<T>: Encodable<S> {
//! fn encode<S:Encoder>(s: &S) {
//! s.emit_rec(|| {
//! s.emit_field("node", 0, || self.node.encode(s));
//! s.emit_field("span", 1, || self.span.encode(s));
//! })
//! }
//! }
//!
//! impl<
//! D: Decoder,
//! T: Decodable<D>
//! > spanned<T>: Decodable<D> {
//! fn decode(d: &D) -> spanned<T> {
//! d.read_rec(|| {
//! {
//! node: d.read_field("node".to_string(), 0, || decode(d)),
//! span: d.read_field("span".to_string(), 1, || decode(d)),
//! }
//! })
//! }
//! }
//! ```
use ast::{MetaItem, Item, Expr, ExprRet, MutMutable, LitNil};
use codemap::Span;

332
src/libsyntax/ext/deriving/generic/mod.rs
View file

@ -8,174 +8,170 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
/*!
Some code that abstracts away much of the boilerplate of writing
`deriving` instances for traits. Among other things it manages getting
access to the fields of the 4 different sorts of structs and enum
variants, as well as creating the method and impl ast instances.
Supported features (fairly exhaustive):
- Methods taking any number of parameters of any type, and returning
any type, other than vectors, bottom and closures.
- Generating `impl`s for types with type parameters and lifetimes
(e.g. `Option<T>`), the parameters are automatically given the
current trait as a bound. (This includes separate type parameters
and lifetimes for methods.)
- Additional bounds on the type parameters, e.g. the `Ord` instance
requires an explicit `PartialEq` bound at the
moment. (`TraitDef.additional_bounds`)
Unsupported: FIXME #6257: calling methods on reference fields,
e.g. deriving Eq/Ord/Clone don't work on `struct A(&int)`,
because of how the auto-dereferencing happens.
The most important thing for implementers is the `Substructure` and
`SubstructureFields` objects. The latter groups 5 possibilities of the
arguments:
- `Struct`, when `Self` is a struct (including tuple structs, e.g
`struct T(int, char)`).
- `EnumMatching`, when `Self` is an enum and all the arguments are the
same variant of the enum (e.g. `Some(1)`, `Some(3)` and `Some(4)`)
- `EnumNonMatching` when `Self` is an enum and the arguments are not
the same variant (e.g. `None`, `Some(1)` and `None`). If
`const_nonmatching` is true, this will contain an empty list.
- `StaticEnum` and `StaticStruct` for static methods, where the type
being derived upon is either an enum or struct respectively. (Any
argument with type Self is just grouped among the non-self
arguments.)
In the first two cases, the values from the corresponding fields in
all the arguments are grouped together. In the `EnumNonMatching` case
this isn't possible (different variants have different fields), so the
fields are grouped by which argument they come from. There are no
fields with values in the static cases, so these are treated entirely
differently.
The non-static cases have `Option<ident>` in several places associated
with field `expr`s. This represents the name of the field it is
associated with. It is only not `None` when the associated field has
an identifier in the source code. For example, the `x`s in the
following snippet
```rust
struct A { x : int }
struct B(int);
enum C {
C0(int),
C1 { x: int }
}
```
The `int`s in `B` and `C0` don't have an identifier, so the
`Option<ident>`s would be `None` for them.
In the static cases, the structure is summarised, either into the just
spans of the fields or a list of spans and the field idents (for tuple
structs and record structs, respectively), or a list of these, for
enums (one for each variant). For empty struct and empty enum
variants, it is represented as a count of 0.
# Examples
The following simplified `PartialEq` is used for in-code examples:
```rust
trait PartialEq {
fn eq(&self, other: &Self);
}
impl PartialEq for int {
fn eq(&self, other: &int) -> bool {
*self == *other
}
}
```
Some examples of the values of `SubstructureFields` follow, using the
above `PartialEq`, `A`, `B` and `C`.
## Structs
When generating the `expr` for the `A` impl, the `SubstructureFields` is
~~~text
Struct(~[FieldInfo {
span: <span of x>
name: Some(<ident of x>),
self_: <expr for &self.x>,
other: ~[<expr for &other.x]
}])
~~~
For the `B` impl, called with `B(a)` and `B(b)`,
~~~text
Struct(~[FieldInfo {
span: <span of `int`>,
name: None,
<expr for &a>
~[<expr for &b>]
}])
~~~
## Enums
When generating the `expr` for a call with `self == C0(a)` and `other
== C0(b)`, the SubstructureFields is
~~~text
EnumMatching(0, <ast::Variant for C0>,
~[FieldInfo {
span: <span of int>
name: None,
self_: <expr for &a>,
other: ~[<expr for &b>]
}])
~~~
For `C1 {x}` and `C1 {x}`,
~~~text
EnumMatching(1, <ast::Variant for C1>,
~[FieldInfo {
span: <span of x>
name: Some(<ident of x>),
self_: <expr for &self.x>,
other: ~[<expr for &other.x>]
}])
~~~
For `C0(a)` and `C1 {x}` ,
~~~text
EnumNonMatching(~[(0, <ast::Variant for B0>,
~[(<span of int>, None, <expr for &a>)]),
(1, <ast::Variant for B1>,
~[(<span of x>, Some(<ident of x>),
<expr for &other.x>)])])
~~~
(and vice versa, but with the order of the outermost list flipped.)
## Static
A static method on the above would result in,
~~~text
StaticStruct(<ast::StructDef of A>, Named(~[(<ident of x>, <span of x>)]))
StaticStruct(<ast::StructDef of B>, Unnamed(~[<span of x>]))
StaticEnum(<ast::EnumDef of C>, ~[(<ident of C0>, <span of C0>, Unnamed(~[<span of int>])),
(<ident of C1>, <span of C1>,
Named(~[(<ident of x>, <span of x>)]))])
~~~
*/
//! Some code that abstracts away much of the boilerplate of writing
//! `deriving` instances for traits. Among other things it manages getting
//! access to the fields of the 4 different sorts of structs and enum
//! variants, as well as creating the method and impl ast instances.
//!
//! Supported features (fairly exhaustive):
//!
//! - Methods taking any number of parameters of any type, and returning
//! any type, other than vectors, bottom and closures.
//! - Generating `impl`s for types with type parameters and lifetimes
//! (e.g. `Option<T>`), the parameters are automatically given the
//! current trait as a bound. (This includes separate type parameters
//! and lifetimes for methods.)
//! - Additional bounds on the type parameters, e.g. the `Ord` instance
//! requires an explicit `PartialEq` bound at the
//! moment. (`TraitDef.additional_bounds`)
//!
//! Unsupported: FIXME #6257: calling methods on reference fields,
//! e.g. deriving Eq/Ord/Clone don't work on `struct A(&int)`,
//! because of how the auto-dereferencing happens.
//!
//! The most important thing for implementers is the `Substructure` and
//! `SubstructureFields` objects. The latter groups 5 possibilities of the
//! arguments:
//!
//! - `Struct`, when `Self` is a struct (including tuple structs, e.g
//! `struct T(int, char)`).
//! - `EnumMatching`, when `Self` is an enum and all the arguments are the
//! same variant of the enum (e.g. `Some(1)`, `Some(3)` and `Some(4)`)
//! - `EnumNonMatching` when `Self` is an enum and the arguments are not
//! the same variant (e.g. `None`, `Some(1)` and `None`). If
//! `const_nonmatching` is true, this will contain an empty list.
//! - `StaticEnum` and `StaticStruct` for static methods, where the type
//! being derived upon is either an enum or struct respectively. (Any
//! argument with type Self is just grouped among the non-self
//! arguments.)
//!
//! In the first two cases, the values from the corresponding fields in
//! all the arguments are grouped together. In the `EnumNonMatching` case
//! this isn't possible (different variants have different fields), so the
//! fields are grouped by which argument they come from. There are no
//! fields with values in the static cases, so these are treated entirely
//! differently.
//!
//! The non-static cases have `Option<ident>` in several places associated
//! with field `expr`s. This represents the name of the field it is
//! associated with. It is only not `None` when the associated field has
//! an identifier in the source code. For example, the `x`s in the
//! following snippet
//!
//! ```rust
//! struct A { x : int }
//!
//! struct B(int);
//!
//! enum C {
//! C0(int),
//! C1 { x: int }
//! }
//! ```
//!
//! The `int`s in `B` and `C0` don't have an identifier, so the
//! `Option<ident>`s would be `None` for them.
//!
//! In the static cases, the structure is summarised, either into the just
//! spans of the fields or a list of spans and the field idents (for tuple
//! structs and record structs, respectively), or a list of these, for
//! enums (one for each variant). For empty struct and empty enum
//! variants, it is represented as a count of 0.
//!
//! # Examples
//!
//! The following simplified `PartialEq` is used for in-code examples:
//!
//! ```rust
//! trait PartialEq {
//! fn eq(&self, other: &Self);
//! }
//! impl PartialEq for int {
//! fn eq(&self, other: &int) -> bool {
//! *self == *other
//! }
//! }
//! ```
//!
//! Some examples of the values of `SubstructureFields` follow, using the
//! above `PartialEq`, `A`, `B` and `C`.
//!
//! ## Structs
//!
//! When generating the `expr` for the `A` impl, the `SubstructureFields` is
//!
//! ~~~text
//! Struct(~[FieldInfo {
//! span: <span of x>
//! name: Some(<ident of x>),
//! self_: <expr for &self.x>,
//! other: ~[<expr for &other.x]
//! }])
//! ~~~
//!
//! For the `B` impl, called with `B(a)` and `B(b)`,
//!
//! ~~~text
//! Struct(~[FieldInfo {
//! span: <span of `int`>,
//! name: None,
//! <expr for &a>
//! ~[<expr for &b>]
//! }])
//! ~~~
//!
//! ## Enums
//!
//! When generating the `expr` for a call with `self == C0(a)` and `other
//! == C0(b)`, the SubstructureFields is
//!
//! ~~~text
//! EnumMatching(0, <ast::Variant for C0>,
//! ~[FieldInfo {
//! span: <span of int>
//! name: None,
//! self_: <expr for &a>,
//! other: ~[<expr for &b>]
//! }])
//! ~~~
//!
//! For `C1 {x}` and `C1 {x}`,
//!
//! ~~~text
//! EnumMatching(1, <ast::Variant for C1>,
//! ~[FieldInfo {
//! span: <span of x>
//! name: Some(<ident of x>),
//! self_: <expr for &self.x>,
//! other: ~[<expr for &other.x>]
//! }])
//! ~~~
//!
//! For `C0(a)` and `C1 {x}` ,
//!
//! ~~~text
//! EnumNonMatching(~[(0, <ast::Variant for B0>,
//! ~[(<span of int>, None, <expr for &a>)]),
//! (1, <ast::Variant for B1>,
//! ~[(<span of x>, Some(<ident of x>),
//! <expr for &other.x>)])])
//! ~~~
//!
//! (and vice versa, but with the order of the outermost list flipped.)
//!
//! ## Static
//!
//! A static method on the above would result in,
//!
//! ~~~text
//! StaticStruct(<ast::StructDef of A>, Named(~[(<ident of x>, <span of x>)]))
//!
//! StaticStruct(<ast::StructDef of B>, Unnamed(~[<span of x>]))
//!
//! StaticEnum(<ast::EnumDef of C>, ~[(<ident of C0>, <span of C0>, Unnamed(~[<span of int>])),
//! (<ident of C1>, <span of C1>,
//! Named(~[(<ident of x>, <span of x>)]))])
//! ~~~
use std::cell::RefCell;
use std::gc::{Gc, GC};

14
src/libsyntax/ext/deriving/generic/ty.rs
View file

@ -25,8 +25,10 @@ use std::gc::Gc;
/// The types of pointers
pub enum PtrTy<'a> {
Send, // ~
Borrowed(Option<&'a str>, ast::Mutability), // &['lifetime] [mut]
/// ~
Send,
/// &'lifetime mut
Borrowed(Option<&'a str>, ast::Mutability),
}
/// A path, e.g. `::std::option::Option::<int>` (global). Has support
@ -83,12 +85,12 @@ impl<'a> Path<'a> {
/// A type. Supports pointers (except for *), Self, and literals
pub enum Ty<'a> {
Self,
// &/Box/ Ty
/// &/Box/ Ty
Ptr(Box<Ty<'a>>, PtrTy<'a>),
// mod::mod::Type<[lifetime], [Params...]>, including a plain type
// parameter, and things like `int`
/// mod::mod::Type<[lifetime], [Params...]>, including a plain type
/// parameter, and things like `int`
Literal(Path<'a>),
// includes nil
/// includes unit
Tuple(Vec<Ty<'a>> )
}

4
src/libsyntax/ext/deriving/show.rs
View file

@ -55,8 +55,8 @@ pub fn expand_deriving_show(cx: &mut ExtCtxt,
trait_def.expand(cx, mitem, item, push)
}
// we construct a format string and then defer to std::fmt, since that
// knows what's up with formatting at so on.
/// We construct a format string and then defer to std::fmt, since that
/// knows what's up with formatting and so on.
fn show_substructure(cx: &mut ExtCtxt, span: Span,
substr: &Substructure) -> Gc<Expr> {
// build `<name>`, `<name>({}, {}, ...)` or `<name> { <field>: {},

20
src/libsyntax/ext/expand.rs
View file

@ -246,11 +246,11 @@ pub fn expand_expr(e: Gc<ast::Expr>, fld: &mut MacroExpander) -> Gc<ast::Expr> {
}
}
// Rename loop label and expand its loop body
//
// The renaming procedure for loop is different in the sense that the loop
// body is in a block enclosed by loop head so the renaming of loop label
// must be propagated to the enclosed context.
/// Rename loop label and expand its loop body
///
/// The renaming procedure for loop is different in the sense that the loop
/// body is in a block enclosed by loop head so the renaming of loop label
/// must be propagated to the enclosed context.
fn expand_loop_block(loop_block: P<Block>,
opt_ident: Option<Ident>,
fld: &mut MacroExpander) -> (P<Block>, Option<Ident>) {
@ -1150,7 +1150,7 @@ mod test {
use super::{pattern_bindings, expand_crate, contains_macro_escape};
use super::{PatIdentFinder, IdentRenamer, PatIdentRenamer};
use ast;
use ast::{Attribute_, AttrOuter, MetaWord};
use ast::{Attribute_, AttrOuter, MetaWord, Name};
use attr;
use codemap;
use codemap::Spanned;
@ -1665,12 +1665,12 @@ foo_module!()
let f_ident = token::str_to_ident("f");
let x_ident = token::str_to_ident("x");
let int_ident = token::str_to_ident("int");
let renames = vec!((x_ident,16));
let renames = vec!((x_ident,Name(16)));
let mut renamer = IdentRenamer{renames: &renames};
let renamed_crate = renamer.fold_crate(the_crate);
let idents = crate_idents(&renamed_crate);
let resolved : Vec<ast::Name> = idents.iter().map(|id| mtwt::resolve(*id)).collect();
assert_eq!(resolved,vec!(f_ident.name,16,int_ident.name,16,16,16));
assert_eq!(resolved,vec!(f_ident.name,Name(16),int_ident.name,Name(16),Name(16),Name(16)));
}
// test the PatIdentRenamer; only PatIdents get renamed
@ -1680,13 +1680,13 @@ foo_module!()
let f_ident = token::str_to_ident("f");
let x_ident = token::str_to_ident("x");
let int_ident = token::str_to_ident("int");
let renames = vec!((x_ident,16));
let renames = vec!((x_ident,Name(16)));
let mut renamer = PatIdentRenamer{renames: &renames};
let renamed_crate = renamer.fold_crate(the_crate);
let idents = crate_idents(&renamed_crate);
let resolved : Vec<ast::Name> = idents.iter().map(|id| mtwt::resolve(*id)).collect();
let x_name = x_ident.name;
assert_eq!(resolved,vec!(f_ident.name,16,int_ident.name,16,x_name,x_name));
assert_eq!(resolved,vec!(f_ident.name,Name(16),int_ident.name,Name(16),x_name,x_name));
}

16
src/libsyntax/ext/format.rs
View file

@ -37,24 +37,24 @@ struct Context<'a, 'b> {
ecx: &'a mut ExtCtxt<'b>,
fmtsp: Span,
// Parsed argument expressions and the types that we've found so far for
// them.
/// Parsed argument expressions and the types that we've found so far for
/// them.
args: Vec<Gc<ast::Expr>>,
arg_types: Vec<Option<ArgumentType>>,
// Parsed named expressions and the types that we've found for them so far.
// Note that we keep a side-array of the ordering of the named arguments
// found to be sure that we can translate them in the same order that they
// were declared in.
/// Parsed named expressions and the types that we've found for them so far.
/// Note that we keep a side-array of the ordering of the named arguments
/// found to be sure that we can translate them in the same order that they
/// were declared in.
names: HashMap<String, Gc<ast::Expr>>,
name_types: HashMap<String, ArgumentType>,
name_ordering: Vec<String>,
// Collection of the compiled `rt::Piece` structures
/// Collection of the compiled `rt::Piece` structures
pieces: Vec<Gc<ast::Expr>>,
name_positions: HashMap<String, uint>,
method_statics: Vec<Gc<ast::Item>>,
// Updated as arguments are consumed or methods are entered
/// Updated as arguments are consumed or methods are entered
nest_level: uint,
next_arg: uint,
}

124
src/libsyntax/ext/mtwt.rs
View file

@ -21,16 +21,16 @@ use std::cell::RefCell;
use std::rc::Rc;
use std::collections::HashMap;
// the SCTable contains a table of SyntaxContext_'s. It
// represents a flattened tree structure, to avoid having
// managed pointers everywhere (that caused an ICE).
// the mark_memo and rename_memo fields are side-tables
// that ensure that adding the same mark to the same context
// gives you back the same context as before. This shouldn't
// change the semantics--everything here is immutable--but
// it should cut down on memory use *a lot*; applying a mark
// to a tree containing 50 identifiers would otherwise generate
// 50 new contexts
/// The SCTable contains a table of SyntaxContext_'s. It
/// represents a flattened tree structure, to avoid having
/// managed pointers everywhere (that caused an ICE).
/// the mark_memo and rename_memo fields are side-tables
/// that ensure that adding the same mark to the same context
/// gives you back the same context as before. This shouldn't
/// change the semantics--everything here is immutable--but
/// it should cut down on memory use *a lot*; applying a mark
/// to a tree containing 50 identifiers would otherwise generate
/// 50 new contexts
pub struct SCTable {
table: RefCell<Vec<SyntaxContext_>>,
mark_memo: RefCell<HashMap<(SyntaxContext,Mrk),SyntaxContext>>,
@ -41,16 +41,16 @@ pub struct SCTable {
pub enum SyntaxContext_ {
EmptyCtxt,
Mark (Mrk,SyntaxContext),
// flattening the name and syntaxcontext into the rename...
// HIDDEN INVARIANTS:
// 1) the first name in a Rename node
// can only be a programmer-supplied name.
// 2) Every Rename node with a given Name in the
// "to" slot must have the same name and context
// in the "from" slot. In essence, they're all
// pointers to a single "rename" event node.
/// flattening the name and syntaxcontext into the rename...
/// HIDDEN INVARIANTS:
/// 1) the first name in a Rename node
/// can only be a programmer-supplied name.
/// 2) Every Rename node with a given Name in the
/// "to" slot must have the same name and context
/// in the "from" slot. In essence, they're all
/// pointers to a single "rename" event node.
Rename (Ident,Name,SyntaxContext),
// actually, IllegalCtxt may not be necessary.
/// actually, IllegalCtxt may not be necessary.
IllegalCtxt
}
@ -62,7 +62,7 @@ pub fn apply_mark(m: Mrk, ctxt: SyntaxContext) -> SyntaxContext {
with_sctable(|table| apply_mark_internal(m, ctxt, table))
}
// Extend a syntax context with a given mark and sctable (explicit memoization)
/// Extend a syntax context with a given mark and sctable (explicit memoization)
fn apply_mark_internal(m: Mrk, ctxt: SyntaxContext, table: &SCTable) -> SyntaxContext {
let key = (ctxt, m);
let new_ctxt = |_: &(SyntaxContext, Mrk)|
@ -77,13 +77,13 @@ pub fn apply_rename(id: Ident, to:Name,
with_sctable(|table| apply_rename_internal(id, to, ctxt, table))
}
// Extend a syntax context with a given rename and sctable (explicit memoization)
/// Extend a syntax context with a given rename and sctable (explicit memoization)
fn apply_rename_internal(id: Ident,
to: Name,
ctxt: SyntaxContext,
table: &SCTable) -> SyntaxContext {
let key = (ctxt,id,to);
let new_ctxt = |_: &(SyntaxContext, Ident, Mrk)|
let key = (ctxt, id, to);
let new_ctxt = |_: &(SyntaxContext, Ident, Name)|
idx_push(&mut *table.table.borrow_mut(), Rename(id, to, ctxt));
*table.rename_memo.borrow_mut().find_or_insert_with(key, new_ctxt)
@ -141,8 +141,8 @@ pub fn clear_tables() {
with_resolve_table_mut(|table| *table = HashMap::new());
}
// Add a value to the end of a vec, return its index
fn idx_push<T>(vec: &mut Vec<T> , val: T) -> u32 {
/// Add a value to the end of a vec, return its index
fn idx_push<T>(vec: &mut Vec<T>, val: T) -> u32 {
vec.push(val);
(vec.len() - 1) as u32
}
@ -173,8 +173,8 @@ fn with_resolve_table_mut<T>(op: |&mut ResolveTable| -> T) -> T {
}
}
// Resolve a syntax object to a name, per MTWT.
// adding memoization to resolve 500+ seconds in resolve for librustc (!)
/// Resolve a syntax object to a name, per MTWT.
/// adding memoization to resolve 500+ seconds in resolve for librustc (!)
fn resolve_internal(id: Ident,
table: &SCTable,
resolve_table: &mut ResolveTable) -> Name {
@ -264,8 +264,8 @@ pub fn outer_mark(ctxt: SyntaxContext) -> Mrk {
})
}
// Push a name... unless it matches the one on top, in which
// case pop and discard (so two of the same marks cancel)
/// Push a name... unless it matches the one on top, in which
/// case pop and discard (so two of the same marks cancel)
fn xor_push(marks: &mut Vec<Mrk>, mark: Mrk) {
if (marks.len() > 0) && (*marks.last().unwrap() == mark) {
marks.pop().unwrap();
@ -301,8 +301,8 @@ mod tests {
assert_eq!(s.clone(), vec!(14));
}
fn id(n: Name, s: SyntaxContext) -> Ident {
Ident {name: n, ctxt: s}
fn id(n: u32, s: SyntaxContext) -> Ident {
Ident {name: Name(n), ctxt: s}
}
// because of the SCTable, I now need a tidy way of
@ -349,12 +349,12 @@ mod tests {
fn test_unfold_refold(){
let mut t = new_sctable_internal();
let test_sc = vec!(M(3),R(id(101,0),14),M(9));
let test_sc = vec!(M(3),R(id(101,0),Name(14)),M(9));
assert_eq!(unfold_test_sc(test_sc.clone(),EMPTY_CTXT,&mut t),4);
{
let table = t.table.borrow();
assert!(*table.get(2) == Mark(9,0));
assert!(*table.get(3) == Rename(id(101,0),14,2));
assert!(*table.get(3) == Rename(id(101,0),Name(14),2));
assert!(*table.get(4) == Mark(3,3));
}
assert_eq!(refold_test_sc(4,&t),test_sc);
@ -381,8 +381,8 @@ mod tests {
#[test]
fn test_marksof () {
let stopname = 242;
let name1 = 243;
let stopname = Name(242);
let name1 = Name(243);
let mut t = new_sctable_internal();
assert_eq!(marksof_internal (EMPTY_CTXT,stopname,&t),Vec::new());
// FIXME #5074: ANF'd to dodge nested calls
@ -396,16 +396,16 @@ mod tests {
assert_eq! (marksof_internal (ans, stopname,&t), vec!(16));}
// rename where stop doesn't match:
{ let chain = vec!(M(9),
R(id(name1,
R(id(name1.uint() as u32,
apply_mark_internal (4, EMPTY_CTXT,&mut t)),
100101102),
Name(100101102)),
M(14));
let ans = unfold_test_sc(chain,EMPTY_CTXT,&mut t);
assert_eq! (marksof_internal (ans, stopname, &t), vec!(9,14));}
// rename where stop does match
{ let name1sc = apply_mark_internal(4, EMPTY_CTXT, &mut t);
let chain = vec!(M(9),
R(id(name1, name1sc),
R(id(name1.uint() as u32, name1sc),
stopname),
M(14));
let ans = unfold_test_sc(chain,EMPTY_CTXT,&mut t);
@ -419,55 +419,55 @@ mod tests {
let mut t = new_sctable_internal();
let mut rt = HashMap::new();
// - ctxt is MT
assert_eq!(resolve_internal(id(a,EMPTY_CTXT),&mut t, &mut rt),a);
assert_eq!(resolve_internal(id(a,EMPTY_CTXT),&mut t, &mut rt),Name(a));
// - simple ignored marks
{ let sc = unfold_marks(vec!(1,2,3),EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),a);}
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),Name(a));}
// - orthogonal rename where names don't match
{ let sc = unfold_test_sc(vec!(R(id(50,EMPTY_CTXT),51),M(12)),EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),a);}
{ let sc = unfold_test_sc(vec!(R(id(50,EMPTY_CTXT),Name(51)),M(12)),EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),Name(a));}
// - rename where names do match, but marks don't
{ let sc1 = apply_mark_internal(1,EMPTY_CTXT,&mut t);
let sc = unfold_test_sc(vec!(R(id(a,sc1),50),
let sc = unfold_test_sc(vec!(R(id(a,sc1),Name(50)),
M(1),
M(2)),
EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), a);}
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), Name(a));}
// - rename where names and marks match
{ let sc1 = unfold_test_sc(vec!(M(1),M(2)),EMPTY_CTXT,&mut t);
let sc = unfold_test_sc(vec!(R(id(a,sc1),50),M(1),M(2)),EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), 50); }
let sc = unfold_test_sc(vec!(R(id(a,sc1),Name(50)),M(1),M(2)),EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), Name(50)); }
// - rename where names and marks match by literal sharing
{ let sc1 = unfold_test_sc(vec!(M(1),M(2)),EMPTY_CTXT,&mut t);
let sc = unfold_test_sc(vec!(R(id(a,sc1),50)),sc1,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), 50); }
let sc = unfold_test_sc(vec!(R(id(a,sc1),Name(50))),sc1,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), Name(50)); }
// - two renames of the same var.. can only happen if you use
// local-expand to prevent the inner binding from being renamed
// during the rename-pass caused by the first:
println!("about to run bad test");
{ let sc = unfold_test_sc(vec!(R(id(a,EMPTY_CTXT),50),
R(id(a,EMPTY_CTXT),51)),
{ let sc = unfold_test_sc(vec!(R(id(a,EMPTY_CTXT),Name(50)),
R(id(a,EMPTY_CTXT),Name(51))),
EMPTY_CTXT,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), 51); }
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt), Name(51)); }
// the simplest double-rename:
{ let a_to_a50 = apply_rename_internal(id(a,EMPTY_CTXT),50,EMPTY_CTXT,&mut t);
let a50_to_a51 = apply_rename_internal(id(a,a_to_a50),51,a_to_a50,&mut t);
assert_eq!(resolve_internal(id(a,a50_to_a51),&mut t, &mut rt),51);
{ let a_to_a50 = apply_rename_internal(id(a,EMPTY_CTXT),Name(50),EMPTY_CTXT,&mut t);
let a50_to_a51 = apply_rename_internal(id(a,a_to_a50),Name(51),a_to_a50,&mut t);
assert_eq!(resolve_internal(id(a,a50_to_a51),&mut t, &mut rt),Name(51));
// mark on the outside doesn't stop rename:
let sc = apply_mark_internal(9,a50_to_a51,&mut t);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),51);
assert_eq!(resolve_internal(id(a,sc),&mut t, &mut rt),Name(51));
// but mark on the inside does:
let a50_to_a51_b = unfold_test_sc(vec!(R(id(a,a_to_a50),51),
let a50_to_a51_b = unfold_test_sc(vec!(R(id(a,a_to_a50),Name(51)),
M(9)),
a_to_a50,
&mut t);
assert_eq!(resolve_internal(id(a,a50_to_a51_b),&mut t, &mut rt),50);}
assert_eq!(resolve_internal(id(a,a50_to_a51_b),&mut t, &mut rt),Name(50));}
}
#[test]
fn mtwt_resolve_test(){
let a = 40;
assert_eq!(resolve(id(a,EMPTY_CTXT)),a);
assert_eq!(resolve(id(a,EMPTY_CTXT)),Name(a));
}
@ -496,10 +496,10 @@ mod tests {
#[test]
fn new_resolves_test() {
let renames = vec!((Ident{name:23,ctxt:EMPTY_CTXT},24),
(Ident{name:29,ctxt:EMPTY_CTXT},29));
let renames = vec!((Ident{name:Name(23),ctxt:EMPTY_CTXT},Name(24)),
(Ident{name:Name(29),ctxt:EMPTY_CTXT},Name(29)));
let new_ctxt1 = apply_renames(&renames,EMPTY_CTXT);
assert_eq!(resolve(Ident{name:23,ctxt:new_ctxt1}),24);
assert_eq!(resolve(Ident{name:29,ctxt:new_ctxt1}),29);
assert_eq!(resolve(Ident{name:Name(23),ctxt:new_ctxt1}),Name(24));
assert_eq!(resolve(Ident{name:Name(29),ctxt:new_ctxt1}),Name(29));
}
}

62
src/libsyntax/ext/quote.rs
View file

@ -363,6 +363,15 @@ fn mk_ident(cx: &ExtCtxt, sp: Span, ident: ast::Ident) -> Gc<ast::Expr> {
vec!(e_str))
}
// Lift a name to the expr that evaluates to that name
fn mk_name(cx: &ExtCtxt, sp: Span, ident: ast::Ident) -> Gc<ast::Expr> {
let e_str = cx.expr_str(sp, token::get_ident(ident));
cx.expr_method_call(sp,
cx.expr_ident(sp, id_ext("ext_cx")),
id_ext("name_of"),
vec!(e_str))
}
fn mk_ast_path(cx: &ExtCtxt, sp: Span, name: &str) -> Gc<ast::Expr> {
let idents = vec!(id_ext("syntax"), id_ext("ast"), id_ext(name));
cx.expr_path(cx.path_global(sp, idents))
@ -401,68 +410,37 @@ fn mk_token(cx: &ExtCtxt, sp: Span, tok: &token::Token) -> Gc<ast::Expr> {
}
LIT_BYTE(i) => {
let e_byte = cx.expr_lit(sp, ast::LitByte(i));
let e_byte = mk_name(cx, sp, i.ident());
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_BYTE"), vec!(e_byte));
}
LIT_CHAR(i) => {
let e_char = cx.expr_lit(sp, ast::LitChar(i));
let e_char = mk_name(cx, sp, i.ident());
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_CHAR"), vec!(e_char));
}
LIT_INT(i, ity) => {
let s_ity = match ity {
ast::TyI => "TyI",
ast::TyI8 => "TyI8",
ast::TyI16 => "TyI16",
ast::TyI32 => "TyI32",
ast::TyI64 => "TyI64"
};
let e_ity = mk_ast_path(cx, sp, s_ity);
let e_i64 = cx.expr_lit(sp, ast::LitInt(i, ast::TyI64));
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_INT"), vec!(e_i64, e_ity));
LIT_INTEGER(i) => {
let e_int = mk_name(cx, sp, i.ident());
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_INTEGER"), vec!(e_int));
}
LIT_UINT(u, uty) => {
let s_uty = match uty {
ast::TyU => "TyU",
ast::TyU8 => "TyU8",
ast::TyU16 => "TyU16",
ast::TyU32 => "TyU32",
ast::TyU64 => "TyU64"
};
let e_uty = mk_ast_path(cx, sp, s_uty);
let e_u64 = cx.expr_lit(sp, ast::LitUint(u, ast::TyU64));
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_UINT"), vec!(e_u64, e_uty));
}
LIT_INT_UNSUFFIXED(i) => {
let e_i64 = cx.expr_lit(sp, ast::LitInt(i, ast::TyI64));
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_INT_UNSUFFIXED"), vec!(e_i64));
}
LIT_FLOAT(fident, fty) => {
let s_fty = match fty {
ast::TyF32 => "TyF32",
ast::TyF64 => "TyF64",
};
let e_fty = mk_ast_path(cx, sp, s_fty);
let e_fident = mk_ident(cx, sp, fident);
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_FLOAT"), vec!(e_fident, e_fty));
LIT_FLOAT(fident) => {
let e_fident = mk_name(cx, sp, fident.ident());
return cx.expr_call(sp, mk_token_path(cx, sp, "LIT_FLOAT"), vec!(e_fident));
}
LIT_STR(ident) => {
return cx.expr_call(sp,
mk_token_path(cx, sp, "LIT_STR"),
vec!(mk_ident(cx, sp, ident)));
vec!(mk_name(cx, sp, ident.ident())));
}
LIT_STR_RAW(ident, n) => {
return cx.expr_call(sp,
mk_token_path(cx, sp, "LIT_STR_RAW"),
vec!(mk_ident(cx, sp, ident), cx.expr_uint(sp, n)));
vec!(mk_name(cx, sp, ident.ident()), cx.expr_uint(sp, n)));
}
IDENT(ident, b) => {
@ -480,7 +458,7 @@ fn mk_token(cx: &ExtCtxt, sp: Span, tok: &token::Token) -> Gc<ast::Expr> {
DOC_COMMENT(ident) => {
return cx.expr_call(sp,
mk_token_path(cx, sp, "DOC_COMMENT"),
vec!(mk_ident(cx, sp, ident)));
vec!(mk_name(cx, sp, ident.ident())));
}
INTERPOLATED(_) => fail!("quote! with interpolated token"),

14
src/libsyntax/ext/source_util.rs
View file

@ -28,7 +28,7 @@ use std::str;
// the column/row/filename of the expression, or they include
// a given file into the current one.
/* line!(): expands to the current line number */
/// line!(): expands to the current line number
pub fn expand_line(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree])
-> Box<base::MacResult> {
base::check_zero_tts(cx, sp, tts, "line!");
@ -49,9 +49,9 @@ pub fn expand_col(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree])
base::MacExpr::new(cx.expr_uint(topmost.call_site, loc.col.to_uint()))
}
/* file!(): expands to the current filename */
/* The filemap (`loc.file`) contains a bunch more information we could spit
* out if we wanted. */
/// file!(): expands to the current filename */
/// The filemap (`loc.file`) contains a bunch more information we could spit
/// out if we wanted.
pub fn expand_file(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree])
-> Box<base::MacResult> {
base::check_zero_tts(cx, sp, tts, "file!");
@ -82,9 +82,9 @@ pub fn expand_mod(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree])
token::intern_and_get_ident(string.as_slice())))
}
// include! : parse the given file as an expr
// This is generally a bad idea because it's going to behave
// unhygienically.
/// include! : parse the given file as an expr
/// This is generally a bad idea because it's going to behave
/// unhygienically.
pub fn expand_include(cx: &mut ExtCtxt, sp: Span, tts: &[ast::TokenTree])
-> Box<base::MacResult> {
let file = match get_single_str_from_tts(cx, sp, tts, "include!") {

175
src/libsyntax/ext/tt/macro_parser.rs
View file

@ -8,7 +8,72 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
// Earley-like parser for macros.
//! 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
//! commits to fully when it hits one in a grammar). This means that there are no
//! completer or predictor rules, and therefore no need to store one column per
//! token: instead, there's a set of current Earley items and a set of next
//! ones. Instead of NTs, we have a special case for Kleene star. The big-O, in
//! pathological cases, is worse than traditional Earley parsing, but it's an
//! easier fit for Macro-by-Example-style rules, and I think the overhead is
//! lower. (In order to prevent the pathological case, we'd need to lazily
//! construct the resulting `NamedMatch`es at the very end. It'd be a pain,
//! and require more memory to keep around old items, but it would also save
//! overhead)
//!
//! Quick intro to how the parser works:
//!
//! A 'position' is a dot in the middle of a matcher, usually represented as a
//! dot. For example `· a $( a )* a b` is a position, as is `a $( · a )* a b`.
//!
//! The parser walks through the input a character at a time, maintaining a list
//! of items consistent with the current position in the input string: `cur_eis`.
//!
//! As it processes them, it fills up `eof_eis` with items that would be valid if
//! the macro invocation is now over, `bb_eis` with items that are waiting on
//! a Rust nonterminal like `$e:expr`, and `next_eis` with items that are waiting
//! on the a particular token. Most of the logic concerns moving the · through the
//! repetitions indicated by Kleene stars. It only advances or calls out to the
//! real Rust parser when no `cur_eis` items remain
//!
//! Example: Start parsing `a a a a b` against [· a $( a )* a b].
//!
//! Remaining input: `a a a a b`
//! next_eis: [· a $( a )* a b]
//!
//! - - - Advance over an `a`. - - -
//!
//! Remaining input: `a a a b`
//! cur: [a · $( a )* a b]
//! Descend/Skip (first item).
//! next: [a $( · a )* a b] [a $( a )* · a b].
//!
//! - - - Advance over an `a`. - - -
//!
//! Remaining input: `a a b`
//! cur: [a $( a · )* a b] next: [a $( a )* a · b]
//! Finish/Repeat (first item)
//! next: [a $( a )* · a b] [a $( · a )* a b] [a $( a )* a · b]
//!
//! - - - Advance over an `a`. - - - (this looks exactly like the last step)
//!
//! Remaining input: `a b`
//! cur: [a $( a · )* a b] next: [a $( a )* a · b]
//! Finish/Repeat (first item)
//! next: [a $( a )* · a b] [a $( · a )* a b] [a $( a )* a · b]
//!
//! - - - Advance over an `a`. - - - (this looks exactly like the last step)
//!
//! Remaining input: `b`
//! cur: [a $( a · )* a b] next: [a $( a )* a · b]
//! Finish/Repeat (first item)
//! next: [a $( a )* · a b] [a $( · a )* a b]
//!
//! - - - Advance over a `b`. - - -
//!
//! Remaining input: ``
//! eof: [a $( a )* a b ·]
use ast;
use ast::{Matcher, MatchTok, MatchSeq, MatchNonterminal, Ident};
@ -25,75 +90,6 @@ use std::rc::Rc;
use std::gc::GC;
use std::collections::HashMap;
/* 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
commits to fully when it hits one in a grammar). This means that there are no
completer or predictor rules, and therefore no need to store one column per
token: instead, there's a set of current Earley items and a set of next
ones. Instead of NTs, we have a special case for Kleene star. The big-O, in
pathological cases, is worse than traditional Earley parsing, but it's an
easier fit for Macro-by-Example-style rules, and I think the overhead is
lower. (In order to prevent the pathological case, we'd need to lazily
construct the resulting `NamedMatch`es at the very end. It'd be a pain,
and require more memory to keep around old items, but it would also save
overhead)*/
/* Quick intro to how the parser works:
A 'position' is a dot in the middle of a matcher, usually represented as a
dot. For example `· a $( a )* a b` is a position, as is `a $( · a )* a b`.
The parser walks through the input a character at a time, maintaining a list
of items consistent with the current position in the input string: `cur_eis`.
As it processes them, it fills up `eof_eis` with items that would be valid if
the macro invocation is now over, `bb_eis` with items that are waiting on
a Rust nonterminal like `$e:expr`, and `next_eis` with items that are waiting
on the a particular token. Most of the logic concerns moving the · through the
repetitions indicated by Kleene stars. It only advances or calls out to the
real Rust parser when no `cur_eis` items remain
Example: Start parsing `a a a a b` against [· a $( a )* a b].
Remaining input: `a a a a b`
next_eis: [· a $( a )* a b]
- - - Advance over an `a`. - - -
Remaining input: `a a a b`
cur: [a · $( a )* a b]
Descend/Skip (first item).
next: [a $( · a )* a b] [a $( a )* · a b].
- - - Advance over an `a`. - - -
Remaining input: `a a b`
cur: [a $( a · )* a b] next: [a $( a )* a · b]
Finish/Repeat (first item)
next: [a $( a )* · a b] [a $( · a )* a b] [a $( a )* a · b]
- - - Advance over an `a`. - - - (this looks exactly like the last step)
Remaining input: `a b`
cur: [a $( a · )* a b] next: [a $( a )* a · b]
Finish/Repeat (first item)
next: [a $( a )* · a b] [a $( · a )* a b] [a $( a )* a · b]
- - - Advance over an `a`. - - - (this looks exactly like the last step)
Remaining input: `b`
cur: [a $( a · )* a b] next: [a $( a )* a · b]
Finish/Repeat (first item)
next: [a $( a )* · a b] [a $( · a )* a b]
- - - Advance over a `b`. - - -
Remaining input: ``
eof: [a $( a )* a b ·]
*/
/* to avoid costly uniqueness checks, we require that `MatchSeq` always has a
nonempty body. */
@ -147,24 +143,24 @@ pub fn initial_matcher_pos(ms: Vec<Matcher> , sep: Option<Token>, lo: BytePos)
}
}
// NamedMatch is a pattern-match result for a single ast::MatchNonterminal:
// so it is associated with a single ident in a parse, and all
// MatchedNonterminal's in the NamedMatch have the same nonterminal type
// (expr, item, etc). All the leaves in a single NamedMatch correspond to a
// single matcher_nonterminal in the ast::Matcher that produced it.
//
// It should probably be renamed, it has more or less exact correspondence to
// ast::match nodes, and the in-memory structure of a particular NamedMatch
// represents the match that occurred when a particular subset of an
// ast::match -- those ast::Matcher nodes leading to a single
// MatchNonterminal -- was applied to a particular token tree.
//
// The width of each MatchedSeq in the NamedMatch, and the identity of the
// MatchedNonterminal's, will depend on the token tree it was applied to: each
// MatchedSeq corresponds to a single MatchSeq in the originating
// ast::Matcher. The depth of the NamedMatch structure will therefore depend
// only on the nesting depth of ast::MatchSeq's in the originating
// ast::Matcher it was derived from.
/// NamedMatch is a pattern-match result for a single ast::MatchNonterminal:
/// so it is associated with a single ident in a parse, and all
/// MatchedNonterminal's in the NamedMatch have the same nonterminal type
/// (expr, item, etc). All the leaves in a single NamedMatch correspond to a
/// single matcher_nonterminal in the ast::Matcher that produced it.
///
/// It should probably be renamed, it has more or less exact correspondence to
/// ast::match nodes, and the in-memory structure of a particular NamedMatch
/// represents the match that occurred when a particular subset of an
/// ast::match -- those ast::Matcher nodes leading to a single
/// MatchNonterminal -- was applied to a particular token tree.
///
/// The width of each MatchedSeq in the NamedMatch, and the identity of the
/// MatchedNonterminal's, will depend on the token tree it was applied to: each
/// MatchedSeq corresponds to a single MatchSeq in the originating
/// ast::Matcher. The depth of the NamedMatch structure will therefore depend
/// only on the nesting depth of ast::MatchSeq's in the originating
/// ast::Matcher it was derived from.
pub enum NamedMatch {
MatchedSeq(Vec<Rc<NamedMatch>>, codemap::Span),
@ -224,7 +220,8 @@ pub fn parse_or_else(sess: &ParseSess,
}
}
// perform a token equality check, ignoring syntax context (that is, an unhygienic comparison)
/// Perform a token equality check, ignoring syntax context (that is, an
/// unhygienic comparison)
pub fn token_name_eq(t1 : &Token, t2 : &Token) -> bool {
match (t1,t2) {
(&token::IDENT(id1,_),&token::IDENT(id2,_))

8
src/libsyntax/ext/tt/macro_rules.rs
View file

@ -119,7 +119,7 @@ impl MacResult for MacroRulesDefiner {
}
}
// Given `lhses` and `rhses`, this is the new macro we create
/// Given `lhses` and `rhses`, this is the new macro we create
fn generic_extension(cx: &ExtCtxt,
sp: Span,
name: Ident,
@ -193,9 +193,9 @@ fn generic_extension(cx: &ExtCtxt,
cx.span_fatal(best_fail_spot, best_fail_msg.as_slice());
}
// this procedure performs the expansion of the
// macro_rules! macro. It parses the RHS and adds
// an extension to the current context.
/// This procedure performs the expansion of the
/// macro_rules! macro. It parses the RHS and adds
/// an extension to the current context.
pub fn add_new_extension(cx: &mut ExtCtxt,
sp: Span,
name: Ident,

12
src/libsyntax/ext/tt/transcribe.rs
View file

@ -32,7 +32,7 @@ struct TtFrame {
#[deriving(Clone)]
pub struct TtReader<'a> {
pub sp_diag: &'a SpanHandler,
// the unzipped tree:
/// the unzipped tree:
stack: Vec<TtFrame>,
/* for MBE-style macro transcription */
interpolations: HashMap<Ident, Rc<NamedMatch>>,
@ -43,9 +43,9 @@ pub struct TtReader<'a> {
pub cur_span: Span,
}
/** This can do Macro-By-Example transcription. On the other hand, if
* `src` contains no `TTSeq`s and `TTNonterminal`s, `interp` can (and
* should) be none. */
/// This can do Macro-By-Example transcription. On the other hand, if
/// `src` contains no `TTSeq`s and `TTNonterminal`s, `interp` can (and
/// should) be none.
pub fn new_tt_reader<'a>(sp_diag: &'a SpanHandler,
interp: Option<HashMap<Ident, Rc<NamedMatch>>>,
src: Vec<ast::TokenTree> )
@ -138,8 +138,8 @@ fn lockstep_iter_size(t: &TokenTree, r: &TtReader) -> LockstepIterSize {
}
}
// return the next token from the TtReader.
// EFFECT: advances the reader's token field
/// Return the next token from the TtReader.
/// EFFECT: advances the reader's token field
pub fn tt_next_token(r: &mut TtReader) -> TokenAndSpan {
// FIXME(pcwalton): Bad copy?
let ret_val = TokenAndSpan {

14
src/libsyntax/lib.rs
View file

@ -8,15 +8,11 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
/*!
The Rust parser and macro expander.
# Note
This API is completely unstable and subject to change.
*/
//! The Rust parser and macro expander.
//!
//! # Note
//!
//! This API is completely unstable and subject to change.
#![crate_id = "syntax#0.11.0"] // NOTE: remove after stage0
#![crate_name = "syntax"]

46
src/libsyntax/parse/attr.rs
View file

@ -18,7 +18,7 @@ use parse::token::INTERPOLATED;
use std::gc::{Gc, GC};
// a parser that can parse attributes.
/// A parser that can parse attributes.
pub trait ParserAttr {
fn parse_outer_attributes(&mut self) -> Vec<ast::Attribute>;
fn parse_attribute(&mut self, permit_inner: bool) -> ast::Attribute;
@ -30,11 +30,11 @@ pub trait ParserAttr {
}
impl<'a> ParserAttr for Parser<'a> {
// Parse attributes that appear before an item
/// Parse attributes that appear before an item
fn parse_outer_attributes(&mut self) -> Vec<ast::Attribute> {
let mut attrs: Vec<ast::Attribute> = Vec::new();
loop {
debug!("parse_outer_attributes: self.token={:?}",
debug!("parse_outer_attributes: self.token={}",
self.token);
match self.token {
token::POUND => {
@ -43,7 +43,7 @@ impl<'a> ParserAttr for Parser<'a> {
token::DOC_COMMENT(s) => {
let attr = ::attr::mk_sugared_doc_attr(
attr::mk_attr_id(),
self.id_to_interned_str(s),
self.id_to_interned_str(s.ident()),
self.span.lo,
self.span.hi
);
@ -59,10 +59,10 @@ impl<'a> ParserAttr for Parser<'a> {
return attrs;
}
// matches attribute = # ! [ meta_item ]
//
// if permit_inner is true, then a leading `!` indicates an inner
// attribute
/// Matches `attribute = # ! [ meta_item ]`
///
/// If permit_inner is true, then a leading `!` indicates an inner
/// attribute
fn parse_attribute(&mut self, permit_inner: bool) -> ast::Attribute {
debug!("parse_attributes: permit_inner={:?} self.token={:?}",
permit_inner, self.token);
@ -114,17 +114,17 @@ impl<'a> ParserAttr for Parser<'a> {
};
}
// Parse attributes that appear after the opening of an item. These should
// be preceded by an exclamation mark, but we accept and warn about one
// terminated by a semicolon. In addition to a vector of inner attributes,
// this function also returns a vector that may contain the first outer
// attribute of the next item (since we can't know whether the attribute
// is an inner attribute of the containing item or an outer attribute of
// the first contained item until we see the semi).
/// Parse attributes that appear after the opening of an item. These should
/// be preceded by an exclamation mark, but we accept and warn about one
/// terminated by a semicolon. In addition to a vector of inner attributes,
/// this function also returns a vector that may contain the first outer
/// attribute of the next item (since we can't know whether the attribute
/// is an inner attribute of the containing item or an outer attribute of
/// the first contained item until we see the semi).
// matches inner_attrs* outer_attr?
// you can make the 'next' field an Option, but the result is going to be
// more useful as a vector.
/// matches inner_attrs* outer_attr?
/// you can make the 'next' field an Option, but the result is going to be
/// more useful as a vector.
fn parse_inner_attrs_and_next(&mut self)
-> (Vec<ast::Attribute> , Vec<ast::Attribute> ) {
let mut inner_attrs: Vec<ast::Attribute> = Vec::new();
@ -139,7 +139,7 @@ impl<'a> ParserAttr for Parser<'a> {
let Span { lo, hi, .. } = self.span;
self.bump();
attr::mk_sugared_doc_attr(attr::mk_attr_id(),
self.id_to_interned_str(s),
self.id_to_interned_str(s.ident()),
lo,
hi)
}
@ -157,9 +157,9 @@ impl<'a> ParserAttr for Parser<'a> {
(inner_attrs, next_outer_attrs)
}
// matches meta_item = IDENT
// | IDENT = lit
// | IDENT meta_seq
/// matches meta_item = IDENT
/// | IDENT = lit
/// | IDENT meta_seq
fn parse_meta_item(&mut self) -> Gc<ast::MetaItem> {
match self.token {
token::INTERPOLATED(token::NtMeta(e)) => {
@ -201,7 +201,7 @@ impl<'a> ParserAttr for Parser<'a> {
}
}
// matches meta_seq = ( COMMASEP(meta_item) )
/// matches meta_seq = ( COMMASEP(meta_item) )
fn parse_meta_seq(&mut self) -> Vec<Gc<ast::MetaItem>> {
self.parse_seq(&token::LPAREN,
&token::RPAREN,

20
src/libsyntax/parse/classify.rs
View file

@ -15,13 +15,13 @@
use ast;
use std::gc::Gc;
// does this expression require a semicolon to be treated
// as a statement? The negation of this: 'can this expression
// be used as a statement without a semicolon' -- is used
// as an early-bail-out in the parser so that, for instance,
// 'if true {...} else {...}
// |x| 5 '
// isn't parsed as (if true {...} else {...} | x) | 5
/// Does this expression require a semicolon to be treated
/// as a statement? The negation of this: 'can this expression
/// be used as a statement without a semicolon' -- is used
/// as an early-bail-out in the parser so that, for instance,
/// if true {...} else {...}
/// |x| 5
/// isn't parsed as (if true {...} else {...} | x) | 5
pub fn expr_requires_semi_to_be_stmt(e: Gc<ast::Expr>) -> bool {
match e.node {
ast::ExprIf(..)
@ -41,9 +41,9 @@ pub fn expr_is_simple_block(e: Gc<ast::Expr>) -> bool {
}
}
// this statement requires a semicolon after it.
// note that in one case (stmt_semi), we've already
// seen the semicolon, and thus don't need another.
/// this statement requires a semicolon after it.
/// note that in one case (stmt_semi), we've already
/// seen the semicolon, and thus don't need another.
pub fn stmt_ends_with_semi(stmt: &ast::Stmt) -> bool {
return match stmt.node {
ast::StmtDecl(d, _) => {

4
src/libsyntax/parse/common.rs
View file

@ -12,8 +12,8 @@
use parse::token;
// SeqSep : a sequence separator (token)
// and whether a trailing separator is allowed.
/// SeqSep : a sequence separator (token)
/// and whether a trailing separator is allowed.
pub struct SeqSep {
pub sep: Option<token::Token>,
pub trailing_sep_allowed: bool

26
src/libsyntax/parse/lexer/comments.rs
View file

@ -13,7 +13,7 @@ use codemap::{BytePos, CharPos, CodeMap, Pos};
use diagnostic;
use parse::lexer::{is_whitespace, Reader};
use parse::lexer::{StringReader, TokenAndSpan};
use parse::lexer::{is_line_non_doc_comment, is_block_non_doc_comment};
use parse::lexer::is_block_doc_comment;
use parse::lexer;
use parse::token;
@ -24,10 +24,14 @@ use std::uint;
#[deriving(Clone, PartialEq)]
pub enum CommentStyle {
Isolated, // No code on either side of each line of the comment
Trailing, // Code exists to the left of the comment
Mixed, // Code before /* foo */ and after the comment
BlankLine, // Just a manual blank line "\n\n", for layout
/// No code on either side of each line of the comment
Isolated,
/// Code exists to the left of the comment
Trailing,
/// Code before /* foo */ and after the comment
Mixed,
/// Just a manual blank line "\n\n", for layout
BlankLine,
}
#[deriving(Clone)]
@ -38,9 +42,9 @@ pub struct Comment {
}
pub fn is_doc_comment(s: &str) -> bool {
(s.starts_with("///") && !is_line_non_doc_comment(s)) ||
(s.starts_with("///") && super::is_doc_comment(s)) ||
s.starts_with("//!") ||
(s.starts_with("/**") && !is_block_non_doc_comment(s)) ||
(s.starts_with("/**") && is_block_doc_comment(s)) ||
s.starts_with("/*!")
}
@ -198,9 +202,9 @@ fn read_line_comments(rdr: &mut StringReader, code_to_the_left: bool,
}
}
// Returns None if the first col chars of s contain a non-whitespace char.
// Otherwise returns Some(k) where k is first char offset after that leading
// whitespace. Note k may be outside bounds of s.
/// Returns None if the first col chars of s contain a non-whitespace char.
/// Otherwise returns Some(k) where k is first char offset after that leading
/// whitespace. Note k may be outside bounds of s.
fn all_whitespace(s: &str, col: CharPos) -> Option<uint> {
let len = s.len();
let mut col = col.to_uint();
@ -256,7 +260,7 @@ fn read_block_comment(rdr: &mut StringReader,
rdr.bump();
rdr.bump();
}
if !is_block_non_doc_comment(curr_line.as_slice()) {
if is_block_doc_comment(curr_line.as_slice()) {
return
}
assert!(!curr_line.as_slice().contains_char('\n'));

813
src/libsyntax/parse/lexer/mod.rs
View file

File diff suppressed because it is too large Load diff

354
src/libsyntax/parse/mod.rs
View file

@ -10,7 +10,6 @@
//! The main parser interface
use ast;
use codemap::{Span, CodeMap, FileMap};
use diagnostic::{SpanHandler, mk_span_handler, default_handler, Auto};
@ -32,7 +31,7 @@ pub mod common;
pub mod classify;
pub mod obsolete;
// info about a parsing session.
/// Info about a parsing session.
pub struct ParseSess {
pub span_diagnostic: SpanHandler, // better be the same as the one in the reader!
/// Used to determine and report recursive mod inclusions
@ -241,14 +240,14 @@ pub fn file_to_filemap(sess: &ParseSess, path: &Path, spanopt: Option<Span>)
unreachable!()
}
// given a session and a string, add the string to
// the session's codemap and return the new filemap
/// Given a session and a string, add the string to
/// the session's codemap and return the new filemap
pub fn string_to_filemap(sess: &ParseSess, source: String, path: String)
-> Rc<FileMap> {
sess.span_diagnostic.cm.new_filemap(path, source)
}
// given a filemap, produce a sequence of token-trees
/// Given a filemap, produce a sequence of token-trees
pub fn filemap_to_tts(sess: &ParseSess, filemap: Rc<FileMap>)
-> Vec<ast::TokenTree> {
// it appears to me that the cfg doesn't matter here... indeed,
@ -259,7 +258,7 @@ pub fn filemap_to_tts(sess: &ParseSess, filemap: Rc<FileMap>)
p1.parse_all_token_trees()
}
// given tts and cfg, produce a parser
/// Given tts and cfg, produce a parser
pub fn tts_to_parser<'a>(sess: &'a ParseSess,
tts: Vec<ast::TokenTree>,
cfg: ast::CrateConfig) -> Parser<'a> {
@ -267,13 +266,354 @@ pub fn tts_to_parser<'a>(sess: &'a ParseSess,
Parser::new(sess, cfg, box trdr)
}
// abort if necessary
/// Abort if necessary
pub fn maybe_aborted<T>(result: T, mut p: Parser) -> T {
p.abort_if_errors();
result
}
/// Parse a string representing a character literal into its final form.
/// Rather than just accepting/rejecting a given literal, unescapes it as
/// well. Can take any slice prefixed by a character escape. Returns the
/// character and the number of characters consumed.
pub fn char_lit(lit: &str) -> (char, int) {
use std::{num, char};
let mut chars = lit.chars();
let c = match (chars.next(), chars.next()) {
(Some(c), None) if c != '\\' => return (c, 1),
(Some('\\'), Some(c)) => match c {
'"' => Some('"'),
'n' => Some('\n'),
'r' => Some('\r'),
't' => Some('\t'),
'\\' => Some('\\'),
'\'' => Some('\''),
'0' => Some('\0'),
_ => { None }
},
_ => fail!("lexer accepted invalid char escape `{}`", lit)
};
match c {
Some(x) => return (x, 2),
None => { }
}
let msg = format!("lexer should have rejected a bad character escape {}", lit);
let msg2 = msg.as_slice();
let esc: |uint| -> Option<(char, int)> = |len|
num::from_str_radix(lit.slice(2, len), 16)
.and_then(char::from_u32)
.map(|x| (x, len as int));
// Unicode escapes
return match lit.as_bytes()[1] as char {
'x' | 'X' => esc(4),
'u' => esc(6),
'U' => esc(10),
_ => None,
}.expect(msg2);
}
/// Parse a string representing a string literal into its final form. Does
/// unescaping.
pub fn str_lit(lit: &str) -> String {
debug!("parse_str_lit: given {}", lit.escape_default());
let mut res = String::with_capacity(lit.len());
// FIXME #8372: This could be a for-loop if it didn't borrow the iterator
let error = |i| format!("lexer should have rejected {} at {}", lit, i);
/// Eat everything up to a non-whitespace
fn eat<'a>(it: &mut ::std::iter::Peekable<(uint, char), ::std::str::CharOffsets<'a>>) {
loop {
match it.peek().map(|x| x.val1()) {
Some(' ') | Some('\n') | Some('\r') | Some('\t') => {
it.next();
},
_ => { break; }
}
}
}
let mut chars = lit.char_indices().peekable();
loop {
match chars.next() {
Some((i, c)) => {
let em = error(i);
match c {
'\\' => {
if chars.peek().expect(em.as_slice()).val1() == '\n' {
eat(&mut chars);
} else if chars.peek().expect(em.as_slice()).val1() == '\r' {
chars.next();
if chars.peek().expect(em.as_slice()).val1() != '\n' {
fail!("lexer accepted bare CR");
}
eat(&mut chars);
} else {
// otherwise, a normal escape
let (c, n) = char_lit(lit.slice_from(i));
for _ in range(0, n - 1) { // we don't need to move past the first \
chars.next();
}
res.push_char(c);
}
},
'\r' => {
if chars.peek().expect(em.as_slice()).val1() != '\n' {
fail!("lexer accepted bare CR");
}
chars.next();
res.push_char('\n');
}
c => res.push_char(c),
}
},
None => break
}
}
res.shrink_to_fit(); // probably not going to do anything, unless there was an escape.
debug!("parse_str_lit: returning {}", res);
res
}
/// Parse a string representing a raw string literal into its final form. The
/// only operation this does is convert embedded CRLF into a single LF.
pub fn raw_str_lit(lit: &str) -> String {
debug!("raw_str_lit: given {}", lit.escape_default());
let mut res = String::with_capacity(lit.len());
// FIXME #8372: This could be a for-loop if it didn't borrow the iterator
let mut chars = lit.chars().peekable();
loop {
match chars.next() {
Some(c) => {
if c == '\r' {
if *chars.peek().unwrap() != '\n' {
fail!("lexer accepted bare CR");
}
chars.next();
res.push_char('\n');
} else {
res.push_char(c);
}
},
None => break
}
}
res.shrink_to_fit();
res
}
pub fn float_lit(s: &str) -> ast::Lit_ {
debug!("float_lit: {}", s);
// FIXME #2252: bounds checking float literals is defered until trans
let s2 = s.chars().filter(|&c| c != '_').collect::<String>();
let s = s2.as_slice();
let mut ty = None;
if s.ends_with("f32") {
ty = Some(ast::TyF32);
} else if s.ends_with("f64") {
ty = Some(ast::TyF64);
}
match ty {
Some(t) => {
ast::LitFloat(token::intern_and_get_ident(s.slice_to(s.len() - t.suffix_len())), t)
},
None => ast::LitFloatUnsuffixed(token::intern_and_get_ident(s))
}
}
/// Parse a string representing a byte literal into its final form. Similar to `char_lit`
pub fn byte_lit(lit: &str) -> (u8, uint) {
let err = |i| format!("lexer accepted invalid byte literal {} step {}", lit, i);
if lit.len() == 1 {
(lit.as_bytes()[0], 1)
} else {
assert!(lit.as_bytes()[0] == b'\\', err(0i));
let b = match lit.as_bytes()[1] {
b'"' => b'"',
b'n' => b'\n',
b'r' => b'\r',
b't' => b'\t',
b'\\' => b'\\',
b'\'' => b'\'',
b'0' => b'\0',
_ => {
match ::std::num::from_str_radix::<u64>(lit.slice(2, 4), 16) {
Some(c) =>
if c > 0xFF {
fail!(err(2))
} else {
return (c as u8, 4)
},
None => fail!(err(3))
}
}
};
return (b, 2);
}
}
pub fn binary_lit(lit: &str) -> Rc<Vec<u8>> {
let mut res = Vec::with_capacity(lit.len());
// FIXME #8372: This could be a for-loop if it didn't borrow the iterator
let error = |i| format!("lexer should have rejected {} at {}", lit, i);
// binary literals *must* be ASCII, but the escapes don't have to be
let mut chars = lit.as_bytes().iter().enumerate().peekable();
loop {
match chars.next() {
Some((i, &c)) => {
if c == b'\\' {
if *chars.peek().expect(error(i).as_slice()).val1() == b'\n' {
loop {
// eat everything up to a non-whitespace
match chars.peek().map(|x| *x.val1()) {
Some(b' ') | Some(b'\n') | Some(b'\r') | Some(b'\t') => {
chars.next();
},
_ => { break; }
}
}
} else {
// otherwise, a normal escape
let (c, n) = byte_lit(lit.slice_from(i));
for _ in range(0, n - 1) { // we don't need to move past the first \
chars.next();
}
res.push(c);
}
} else {
res.push(c);
}
},
None => { break; }
}
}
Rc::new(res)
}
pub fn integer_lit(s: &str, sd: &SpanHandler, sp: Span) -> ast::Lit_ {
// s can only be ascii, byte indexing is fine
let s2 = s.chars().filter(|&c| c != '_').collect::<String>();
let mut s = s2.as_slice();
debug!("parse_integer_lit: {}", s);
if s.len() == 1 {
return ast::LitIntUnsuffixed((s.char_at(0)).to_digit(10).unwrap() as i64);
}
let mut base = 10;
let orig = s;
#[deriving(Show)]
enum Result {
Nothing,
Signed(ast::IntTy),
Unsigned(ast::UintTy)
}
impl Result {
fn suffix_len(&self) -> uint {
match *self {
Nothing => 0,
Signed(s) => s.suffix_len(),
Unsigned(u) => u.suffix_len()
}
}
}
let mut ty = Nothing;
if s.char_at(0) == '0' {
match s.char_at(1) {
'x' => base = 16,
'o' => base = 8,
'b' => base = 2,
_ => { }
}
}
if base != 10 {
s = s.slice_from(2);
}
let last = s.len() - 1;
match s.char_at(last) {
'i' => ty = Signed(ast::TyI),
'u' => ty = Unsigned(ast::TyU),
'8' => {
if s.len() > 2 {
match s.char_at(last - 1) {
'i' => ty = Signed(ast::TyI8),
'u' => ty = Unsigned(ast::TyU8),
_ => { }
}
}
},
'6' => {
if s.len() > 3 && s.char_at(last - 1) == '1' {
match s.char_at(last - 2) {
'i' => ty = Signed(ast::TyI16),
'u' => ty = Unsigned(ast::TyU16),
_ => { }
}
}
},
'2' => {
if s.len() > 3 && s.char_at(last - 1) == '3' {
match s.char_at(last - 2) {
'i' => ty = Signed(ast::TyI32),
'u' => ty = Unsigned(ast::TyU32),
_ => { }
}
}
},
'4' => {
if s.len() > 3 && s.char_at(last - 1) == '6' {
match s.char_at(last - 2) {
'i' => ty = Signed(ast::TyI64),
'u' => ty = Unsigned(ast::TyU64),
_ => { }
}
}
},
_ => { }
}
s = s.slice_to(s.len() - ty.suffix_len());
debug!("The suffix is {}, base {}, the new string is {}, the original \
string was {}", ty, base, s, orig);
let res: u64 = match ::std::num::from_str_radix(s, base) {
Some(r) => r,
None => { sd.span_err(sp, "int literal is too large"); 0 }
};
match ty {
Nothing => ast::LitIntUnsuffixed(res as i64),
Signed(t) => ast::LitInt(res as i64, t),
Unsigned(t) => ast::LitUint(res, t)
}
}
#[cfg(test)]
mod test {

8
src/libsyntax/parse/obsolete.rs
View file

@ -38,8 +38,8 @@ pub enum ObsoleteSyntax {
pub trait ParserObsoleteMethods {
/// Reports an obsolete syntax non-fatal error.
fn obsolete(&mut self, sp: Span, kind: ObsoleteSyntax);
// Reports an obsolete syntax non-fatal error, and returns
// a placeholder expression
/// Reports an obsolete syntax non-fatal error, and returns
/// a placeholder expression
fn obsolete_expr(&mut self, sp: Span, kind: ObsoleteSyntax) -> Gc<Expr>;
fn report(&mut self,
sp: Span,
@ -83,8 +83,8 @@ impl<'a> ParserObsoleteMethods for parser::Parser<'a> {
self.report(sp, kind, kind_str, desc);
}
// Reports an obsolete syntax non-fatal error, and returns
// a placeholder expression
/// Reports an obsolete syntax non-fatal error, and returns
/// a placeholder expression
fn obsolete_expr(&mut self, sp: Span, kind: ObsoleteSyntax) -> Gc<Expr> {
self.obsolete(sp, kind);
self.mk_expr(sp.lo, sp.hi, ExprLit(box(GC) respan(sp, LitNil)))

517
src/libsyntax/parse/parser.rs
View file

File diff suppressed because it is too large Load diff

196
src/libsyntax/parse/token.rs
View file

@ -10,7 +10,6 @@
use ast;
use ast::{P, Ident, Name, Mrk};
use ast_util;
use ext::mtwt;
use parse::token;
use util::interner::{RcStr, StrInterner};
@ -79,30 +78,37 @@ pub enum Token {
QUESTION,
/* Literals */
LIT_BYTE(u8),
LIT_CHAR(char),
LIT_INT(i64, ast::IntTy),
LIT_UINT(u64, ast::UintTy),
LIT_INT_UNSUFFIXED(i64),
LIT_FLOAT(ast::Ident, ast::FloatTy),
LIT_FLOAT_UNSUFFIXED(ast::Ident),
LIT_STR(ast::Ident),
LIT_STR_RAW(ast::Ident, uint), /* raw str delimited by n hash symbols */
LIT_BINARY(Rc<Vec<u8>>),
LIT_BINARY_RAW(Rc<Vec<u8>>, uint), /* raw binary str delimited by n hash symbols */
LIT_BYTE(Name),
LIT_CHAR(Name),
LIT_INTEGER(Name),
LIT_FLOAT(Name),
LIT_STR(Name),
LIT_STR_RAW(Name, uint), /* raw str delimited by n hash symbols */
LIT_BINARY(Name),
LIT_BINARY_RAW(Name, uint), /* raw binary str delimited by n hash symbols */
/* Name components */
// an identifier contains an "is_mod_name" boolean,
// indicating whether :: follows this token with no
// whitespace in between.
IDENT(ast::Ident, bool),
/// An identifier contains an "is_mod_name" boolean,
/// indicating whether :: follows this token with no
/// whitespace in between.
IDENT(Ident, bool),
UNDERSCORE,
LIFETIME(ast::Ident),
LIFETIME(Ident),
/* For interpolation */
INTERPOLATED(Nonterminal),
DOC_COMMENT(Name),
// Junk. These carry no data because we don't really care about the data
// they *would* carry, and don't really want to allocate a new ident for
// them. Instead, users could extract that from the associated span.
/// Whitespace
WS,
/// Comment
COMMENT,
SHEBANG(Name),
DOC_COMMENT(ast::Ident),
EOF,
}
@ -115,11 +121,12 @@ pub enum Nonterminal {
NtPat( Gc<ast::Pat>),
NtExpr(Gc<ast::Expr>),
NtTy( P<ast::Ty>),
// see IDENT, above, for meaning of bool in NtIdent:
NtIdent(Box<ast::Ident>, bool),
NtMeta(Gc<ast::MetaItem>), // stuff inside brackets for attributes
/// See IDENT, above, for meaning of bool in NtIdent:
NtIdent(Box<Ident>, bool),
/// Stuff inside brackets for attributes
NtMeta(Gc<ast::MetaItem>),
NtPath(Box<ast::Path>),
NtTT( Gc<ast::TokenTree>), // needs @ed to break a circularity
NtTT( Gc<ast::TokenTree>), // needs Gc'd to break a circularity
NtMatchers(Vec<ast::Matcher> )
}
@ -200,54 +207,28 @@ pub fn to_string(t: &Token) -> String {
/* Literals */
LIT_BYTE(b) => {
let mut res = String::from_str("b'");
(b as char).escape_default(|c| {
res.push_char(c);
});
res.push_char('\'');
res
format!("b'{}'", b.as_str())
}
LIT_CHAR(c) => {
let mut res = String::from_str("'");
c.escape_default(|c| {
res.push_char(c);
});
res.push_char('\'');
res
format!("'{}'", c.as_str())
}
LIT_INT(i, t) => ast_util::int_ty_to_string(t, Some(i)),
LIT_UINT(u, t) => ast_util::uint_ty_to_string(t, Some(u)),
LIT_INT_UNSUFFIXED(i) => { (i as u64).to_string() }
LIT_FLOAT(s, t) => {
let mut body = String::from_str(get_ident(s).get());
if body.as_slice().ends_with(".") {
body.push_char('0'); // `10.f` is not a float literal
}
body.push_str(ast_util::float_ty_to_string(t).as_slice());
body
}
LIT_FLOAT_UNSUFFIXED(s) => {
let mut body = String::from_str(get_ident(s).get());
if body.as_slice().ends_with(".") {
body.push_char('0'); // `10.f` is not a float literal
}
body
LIT_INTEGER(c) | LIT_FLOAT(c) => {
c.as_str().to_string()
}
LIT_STR(s) => {
format!("\"{}\"", get_ident(s).get().escape_default())
format!("\"{}\"", s.as_str())
}
LIT_STR_RAW(s, n) => {
format!("r{delim}\"{string}\"{delim}",
delim="#".repeat(n), string=get_ident(s))
delim="#".repeat(n), string=s.as_str())
}
LIT_BINARY(ref v) => {
format!(
"b\"{}\"",
v.iter().map(|&b| b as char).collect::<String>().escape_default())
LIT_BINARY(v) => {
format!("b\"{}\"", v.as_str())
}
LIT_BINARY_RAW(ref s, n) => {
LIT_BINARY_RAW(s, n) => {
format!("br{delim}\"{string}\"{delim}",
delim="#".repeat(n), string=s.as_slice().to_ascii().as_str_ascii())
delim="#".repeat(n), string=s.as_str())
}
/* Name components */
@ -258,8 +239,12 @@ pub fn to_string(t: &Token) -> String {
UNDERSCORE => "_".to_string(),
/* Other */
DOC_COMMENT(s) => get_ident(s).get().to_string(),
DOC_COMMENT(s) => s.as_str().to_string(),
EOF => "<eof>".to_string(),
WS => " ".to_string(),
COMMENT => "/* */".to_string(),
SHEBANG(s) => format!("/* shebang: {}*/", s.as_str()),
INTERPOLATED(ref nt) => {
match nt {
&NtExpr(ref e) => ::print::pprust::expr_to_string(&**e),
@ -296,11 +281,8 @@ pub fn can_begin_expr(t: &Token) -> bool {
TILDE => true,
LIT_BYTE(_) => true,
LIT_CHAR(_) => true,
LIT_INT(_, _) => true,
LIT_UINT(_, _) => true,
LIT_INT_UNSUFFIXED(_) => true,
LIT_FLOAT(_, _) => true,
LIT_FLOAT_UNSUFFIXED(_) => true,
LIT_INTEGER(_) => true,
LIT_FLOAT(_) => true,
LIT_STR(_) => true,
LIT_STR_RAW(_, _) => true,
LIT_BINARY(_) => true,
@ -337,11 +319,8 @@ pub fn is_lit(t: &Token) -> bool {
match *t {
LIT_BYTE(_) => true,
LIT_CHAR(_) => true,
LIT_INT(_, _) => true,
LIT_UINT(_, _) => true,
LIT_INT_UNSUFFIXED(_) => true,
LIT_FLOAT(_, _) => true,
LIT_FLOAT_UNSUFFIXED(_) => true,
LIT_INTEGER(_) => true,
LIT_FLOAT(_) => true,
LIT_STR(_) => true,
LIT_STR_RAW(_, _) => true,
LIT_BINARY(_) => true,
@ -395,19 +374,19 @@ macro_rules! declare_special_idents_and_keywords {(
$( ($rk_name:expr, $rk_variant:ident, $rk_str:expr); )*
}
) => {
static STRICT_KEYWORD_START: Name = first!($( $sk_name, )*);
static STRICT_KEYWORD_FINAL: Name = last!($( $sk_name, )*);
static RESERVED_KEYWORD_START: Name = first!($( $rk_name, )*);
static RESERVED_KEYWORD_FINAL: Name = last!($( $rk_name, )*);
static STRICT_KEYWORD_START: Name = first!($( Name($sk_name), )*);
static STRICT_KEYWORD_FINAL: Name = last!($( Name($sk_name), )*);
static RESERVED_KEYWORD_START: Name = first!($( Name($rk_name), )*);
static RESERVED_KEYWORD_FINAL: Name = last!($( Name($rk_name), )*);
pub mod special_idents {
use ast::Ident;
$( pub static $si_static: Ident = Ident { name: $si_name, ctxt: 0 }; )*
use ast::{Ident, Name};
$( pub static $si_static: Ident = Ident { name: Name($si_name), ctxt: 0 }; )*
}
pub mod special_names {
use ast::Name;
$( pub static $si_static: Name = $si_name; )*
$( pub static $si_static: Name = Name($si_name); )*
}
/**
@ -428,8 +407,8 @@ macro_rules! declare_special_idents_and_keywords {(
impl Keyword {
pub fn to_name(&self) -> Name {
match *self {
$( $sk_variant => $sk_name, )*
$( $rk_variant => $rk_name, )*
$( $sk_variant => Name($sk_name), )*
$( $rk_variant => Name($rk_name), )*
}
}
}
@ -448,8 +427,11 @@ macro_rules! declare_special_idents_and_keywords {(
}}
// If the special idents get renumbered, remember to modify these two as appropriate
pub static SELF_KEYWORD_NAME: Name = 1;
static STATIC_KEYWORD_NAME: Name = 2;
pub static SELF_KEYWORD_NAME: Name = Name(SELF_KEYWORD_NAME_NUM);
static STATIC_KEYWORD_NAME: Name = Name(STATIC_KEYWORD_NAME_NUM);
pub static SELF_KEYWORD_NAME_NUM: u32 = 1;
static STATIC_KEYWORD_NAME_NUM: u32 = 2;
// NB: leaving holes in the ident table is bad! a different ident will get
// interned with the id from the hole, but it will be between the min and max
@ -459,8 +441,8 @@ declare_special_idents_and_keywords! {
pub mod special_idents {
// These ones are statics
(0, invalid, "");
(super::SELF_KEYWORD_NAME, self_, "self");
(super::STATIC_KEYWORD_NAME, statik, "static");
(super::SELF_KEYWORD_NAME_NUM, self_, "self");
(super::STATIC_KEYWORD_NAME_NUM, statik, "static");
(3, static_lifetime, "'static");
// for matcher NTs
@ -500,8 +482,8 @@ declare_special_idents_and_keywords! {
(29, Ref, "ref");
(30, Return, "return");
// Static and Self are also special idents (prefill de-dupes)
(super::STATIC_KEYWORD_NAME, Static, "static");
(super::SELF_KEYWORD_NAME, Self, "self");
(super::STATIC_KEYWORD_NAME_NUM, Static, "static");
(super::SELF_KEYWORD_NAME_NUM, Self, "self");
(31, Struct, "struct");
(32, Super, "super");
(33, True, "true");
@ -683,20 +665,20 @@ pub fn gensym(s: &str) -> Name {
/// Maps a string to an identifier with an empty syntax context.
#[inline]
pub fn str_to_ident(s: &str) -> ast::Ident {
ast::Ident::new(intern(s))
pub fn str_to_ident(s: &str) -> Ident {
Ident::new(intern(s))
}
/// Maps a string to a gensym'ed identifier.
#[inline]
pub fn gensym_ident(s: &str) -> ast::Ident {
ast::Ident::new(gensym(s))
pub fn gensym_ident(s: &str) -> Ident {
Ident::new(gensym(s))
}
// create a fresh name that maps to the same string as the old one.
// note that this guarantees that str_ptr_eq(ident_to_string(src),interner_get(fresh_name(src)));
// that is, that the new name and the old one are connected to ptr_eq strings.
pub fn fresh_name(src: &ast::Ident) -> Name {
pub fn fresh_name(src: &Ident) -> Name {
let interner = get_ident_interner();
interner.gensym_copy(src.name)
// following: debug version. Could work in final except that it's incompatible with
@ -708,7 +690,7 @@ pub fn fresh_name(src: &ast::Ident) -> Name {
// create a fresh mark.
pub fn fresh_mark() -> Mrk {
gensym("mark")
gensym("mark").uint() as u32
}
// See the macro above about the types of keywords
@ -722,10 +704,13 @@ pub fn is_keyword(kw: keywords::Keyword, tok: &Token) -> bool {
pub fn is_any_keyword(tok: &Token) -> bool {
match *tok {
token::IDENT(sid, false) => match sid.name {
SELF_KEYWORD_NAME | STATIC_KEYWORD_NAME |
STRICT_KEYWORD_START .. RESERVED_KEYWORD_FINAL => true,
_ => false,
token::IDENT(sid, false) => {
let n = sid.name;
n == SELF_KEYWORD_NAME
|| n == STATIC_KEYWORD_NAME
|| STRICT_KEYWORD_START <= n
&& n <= RESERVED_KEYWORD_FINAL
},
_ => false
}
@ -733,10 +718,13 @@ pub fn is_any_keyword(tok: &Token) -> bool {
pub fn is_strict_keyword(tok: &Token) -> bool {
match *tok {
token::IDENT(sid, false) => match sid.name {
SELF_KEYWORD_NAME | STATIC_KEYWORD_NAME |
STRICT_KEYWORD_START .. STRICT_KEYWORD_FINAL => true,
_ => false,
token::IDENT(sid, false) => {
let n = sid.name;
n == SELF_KEYWORD_NAME
|| n == STATIC_KEYWORD_NAME
|| STRICT_KEYWORD_START <= n
&& n <= STRICT_KEYWORD_FINAL
},
_ => false,
}
@ -744,9 +732,11 @@ pub fn is_strict_keyword(tok: &Token) -> bool {
pub fn is_reserved_keyword(tok: &Token) -> bool {
match *tok {
token::IDENT(sid, false) => match sid.name {
RESERVED_KEYWORD_START .. RESERVED_KEYWORD_FINAL => true,
_ => false,
token::IDENT(sid, false) => {
let n = sid.name;
RESERVED_KEYWORD_START <= n
&& n <= RESERVED_KEYWORD_FINAL
},
_ => false,
}
@ -768,7 +758,7 @@ mod test {
use ext::mtwt;
fn mark_ident(id : ast::Ident, m : ast::Mrk) -> ast::Ident {
ast::Ident{name:id.name,ctxt:mtwt::apply_mark(m,id.ctxt)}
ast::Ident { name: id.name, ctxt:mtwt::apply_mark(m, id.ctxt) }
}
#[test] fn mtwt_token_eq_test() {

303
src/libsyntax/print/pp.rs
View file

@ -8,58 +8,56 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
/*
* This pretty-printer is a direct reimplementation of Philip Karlton's
* Mesa pretty-printer, as described in appendix A of
*
* STAN-CS-79-770: "Pretty Printing", by Derek C. Oppen.
* Stanford Department of Computer Science, 1979.
*
* The algorithm's aim is to break a stream into as few lines as possible
* while respecting the indentation-consistency requirements of the enclosing
* block, and avoiding breaking at silly places on block boundaries, for
* example, between "x" and ")" in "x)".
*
* I am implementing this algorithm because it comes with 20 pages of
* documentation explaining its theory, and because it addresses the set of
* concerns I've seen other pretty-printers fall down on. Weirdly. Even though
* it's 32 years old. What can I say?
*
* Despite some redundancies and quirks in the way it's implemented in that
* paper, I've opted to keep the implementation here as similar as I can,
* changing only what was blatantly wrong, a typo, or sufficiently
* non-idiomatic rust that it really stuck out.
*
* In particular you'll see a certain amount of churn related to INTEGER vs.
* CARDINAL in the Mesa implementation. Mesa apparently interconverts the two
* somewhat readily? In any case, I've used uint for indices-in-buffers and
* ints for character-sizes-and-indentation-offsets. This respects the need
* for ints to "go negative" while carrying a pending-calculation balance, and
* helps differentiate all the numbers flying around internally (slightly).
*
* I also inverted the indentation arithmetic used in the print stack, since
* the Mesa implementation (somewhat randomly) stores the offset on the print
* stack in terms of margin-col rather than col itself. I store col.
*
* I also implemented a small change in the String token, in that I store an
* explicit length for the string. For most tokens this is just the length of
* the accompanying string. But it's necessary to permit it to differ, for
* encoding things that are supposed to "go on their own line" -- certain
* classes of comment and blank-line -- where relying on adjacent
* hardbreak-like Break tokens with long blankness indication doesn't actually
* work. To see why, consider when there is a "thing that should be on its own
* line" between two long blocks, say functions. If you put a hardbreak after
* each function (or before each) and the breaking algorithm decides to break
* there anyways (because the functions themselves are long) you wind up with
* extra blank lines. If you don't put hardbreaks you can wind up with the
* "thing which should be on its own line" not getting its own line in the
* rare case of "really small functions" or such. This re-occurs with comments
* and explicit blank lines. So in those cases we use a string with a payload
* we want isolated to a line and an explicit length that's huge, surrounded
* by two zero-length breaks. The algorithm will try its best to fit it on a
* line (which it can't) and so naturally place the content on its own line to
* avoid combining it with other lines and making matters even worse.
*/
//! This pretty-printer is a direct reimplementation of Philip Karlton's
//! Mesa pretty-printer, as described in appendix A of
//!
//! STAN-CS-79-770: "Pretty Printing", by Derek C. Oppen.
//! Stanford Department of Computer Science, 1979.
//!
//! The algorithm's aim is to break a stream into as few lines as possible
//! while respecting the indentation-consistency requirements of the enclosing
//! block, and avoiding breaking at silly places on block boundaries, for
//! example, between "x" and ")" in "x)".
//!
//! I am implementing this algorithm because it comes with 20 pages of
//! documentation explaining its theory, and because it addresses the set of
//! concerns I've seen other pretty-printers fall down on. Weirdly. Even though
//! it's 32 years old. What can I say?
//!
//! Despite some redundancies and quirks in the way it's implemented in that
//! paper, I've opted to keep the implementation here as similar as I can,
//! changing only what was blatantly wrong, a typo, or sufficiently
//! non-idiomatic rust that it really stuck out.
//!
//! In particular you'll see a certain amount of churn related to INTEGER vs.
//! CARDINAL in the Mesa implementation. Mesa apparently interconverts the two
//! somewhat readily? In any case, I've used uint for indices-in-buffers and
//! ints for character-sizes-and-indentation-offsets. This respects the need
//! for ints to "go negative" while carrying a pending-calculation balance, and
//! helps differentiate all the numbers flying around internally (slightly).
//!
//! I also inverted the indentation arithmetic used in the print stack, since
//! the Mesa implementation (somewhat randomly) stores the offset on the print
//! stack in terms of margin-col rather than col itself. I store col.
//!
//! I also implemented a small change in the String token, in that I store an
//! explicit length for the string. For most tokens this is just the length of
//! the accompanying string. But it's necessary to permit it to differ, for
//! encoding things that are supposed to "go on their own line" -- certain
//! classes of comment and blank-line -- where relying on adjacent
//! hardbreak-like Break tokens with long blankness indication doesn't actually
//! work. To see why, consider when there is a "thing that should be on its own
//! line" between two long blocks, say functions. If you put a hardbreak after
//! each function (or before each) and the breaking algorithm decides to break
//! there anyways (because the functions themselves are long) you wind up with
//! extra blank lines. If you don't put hardbreaks you can wind up with the
//! "thing which should be on its own line" not getting its own line in the
//! rare case of "really small functions" or such. This re-occurs with comments
//! and explicit blank lines. So in those cases we use a string with a payload
//! we want isolated to a line and an explicit length that's huge, surrounded
//! by two zero-length breaks. The algorithm will try its best to fit it on a
//! line (which it can't) and so naturally place the content on its own line to
//! avoid combining it with other lines and making matters even worse.
use std::io;
use std::string::String;
@ -186,107 +184,116 @@ pub fn mk_printer(out: Box<io::Writer>, linewidth: uint) -> Printer {
}
/*
* In case you do not have the paper, here is an explanation of what's going
* on.
*
* There is a stream of input tokens flowing through this printer.
*
* The printer buffers up to 3N tokens inside itself, where N is linewidth.
* Yes, linewidth is chars and tokens are multi-char, but in the worst
* case every token worth buffering is 1 char long, so it's ok.
*
* Tokens are String, Break, and Begin/End to delimit blocks.
*
* Begin tokens can carry an offset, saying "how far to indent when you break
* inside here", as well as a flag indicating "consistent" or "inconsistent"
* breaking. Consistent breaking means that after the first break, no attempt
* will be made to flow subsequent breaks together onto lines. Inconsistent
* is the opposite. Inconsistent breaking example would be, say:
*
* foo(hello, there, good, friends)
*
* breaking inconsistently to become
*
* foo(hello, there
* good, friends);
*
* whereas a consistent breaking would yield:
*
* foo(hello,
* there
* good,
* friends);
*
* That is, in the consistent-break blocks we value vertical alignment
* more than the ability to cram stuff onto a line. But in all cases if it
* can make a block a one-liner, it'll do so.
*
* Carrying on with high-level logic:
*
* The buffered tokens go through a ring-buffer, 'tokens'. The 'left' and
* 'right' indices denote the active portion of the ring buffer as well as
* describing hypothetical points-in-the-infinite-stream at most 3N tokens
* apart (i.e. "not wrapped to ring-buffer boundaries"). The paper will switch
* between using 'left' and 'right' terms to denote the wrapepd-to-ring-buffer
* and point-in-infinite-stream senses freely.
*
* There is a parallel ring buffer, 'size', that holds the calculated size of
* each token. Why calculated? Because for Begin/End pairs, the "size"
* includes everything between the pair. That is, the "size" of Begin is
* actually the sum of the sizes of everything between Begin and the paired
* End that follows. Since that is arbitrarily far in the future, 'size' is
* being rewritten regularly while the printer runs; in fact most of the
* machinery is here to work out 'size' entries on the fly (and give up when
* they're so obviously over-long that "infinity" is a good enough
* approximation for purposes of line breaking).
*
* The "input side" of the printer is managed as an abstract process called
* SCAN, which uses 'scan_stack', 'scan_stack_empty', 'top' and 'bottom', to
* manage calculating 'size'. SCAN is, in other words, the process of
* calculating 'size' entries.
*
* The "output side" of the printer is managed by an abstract process called
* PRINT, which uses 'print_stack', 'margin' and 'space' to figure out what to
* do with each token/size pair it consumes as it goes. It's trying to consume
* the entire buffered window, but can't output anything until the size is >=
* 0 (sizes are set to negative while they're pending calculation).
*
* So SCAN takes input and buffers tokens and pending calculations, while
* PRINT gobbles up completed calculations and tokens from the buffer. The
* theory is that the two can never get more than 3N tokens apart, because
* once there's "obviously" too much data to fit on a line, in a size
* calculation, SCAN will write "infinity" to the size and let PRINT consume
* it.
*
* In this implementation (following the paper, again) the SCAN process is
* the method called 'pretty_print', and the 'PRINT' process is the method
* called 'print'.
*/
/// In case you do not have the paper, here is an explanation of what's going
/// on.
///
/// There is a stream of input tokens flowing through this printer.
///
/// The printer buffers up to 3N tokens inside itself, where N is linewidth.
/// Yes, linewidth is chars and tokens are multi-char, but in the worst
/// case every token worth buffering is 1 char long, so it's ok.
///
/// Tokens are String, Break, and Begin/End to delimit blocks.
///
/// Begin tokens can carry an offset, saying "how far to indent when you break
/// inside here", as well as a flag indicating "consistent" or "inconsistent"
/// breaking. Consistent breaking means that after the first break, no attempt
/// will be made to flow subsequent breaks together onto lines. Inconsistent
/// is the opposite. Inconsistent breaking example would be, say:
///
/// foo(hello, there, good, friends)
///
/// breaking inconsistently to become
///
/// foo(hello, there
/// good, friends);
///
/// whereas a consistent breaking would yield:
///
/// foo(hello,
/// there
/// good,
/// friends);
///
/// That is, in the consistent-break blocks we value vertical alignment
/// more than the ability to cram stuff onto a line. But in all cases if it
/// can make a block a one-liner, it'll do so.
///
/// Carrying on with high-level logic:
///
/// The buffered tokens go through a ring-buffer, 'tokens'. The 'left' and
/// 'right' indices denote the active portion of the ring buffer as well as
/// describing hypothetical points-in-the-infinite-stream at most 3N tokens
/// apart (i.e. "not wrapped to ring-buffer boundaries"). The paper will switch
/// between using 'left' and 'right' terms to denote the wrapepd-to-ring-buffer
/// and point-in-infinite-stream senses freely.
///
/// There is a parallel ring buffer, 'size', that holds the calculated size of
/// each token. Why calculated? Because for Begin/End pairs, the "size"
/// includes everything betwen the pair. That is, the "size" of Begin is
/// actually the sum of the sizes of everything between Begin and the paired
/// End that follows. Since that is arbitrarily far in the future, 'size' is
/// being rewritten regularly while the printer runs; in fact most of the
/// machinery is here to work out 'size' entries on the fly (and give up when
/// they're so obviously over-long that "infinity" is a good enough
/// approximation for purposes of line breaking).
///
/// The "input side" of the printer is managed as an abstract process called
/// SCAN, which uses 'scan_stack', 'scan_stack_empty', 'top' and 'bottom', to
/// manage calculating 'size'. SCAN is, in other words, the process of
/// calculating 'size' entries.
///
/// The "output side" of the printer is managed by an abstract process called
/// PRINT, which uses 'print_stack', 'margin' and 'space' to figure out what to
/// do with each token/size pair it consumes as it goes. It's trying to consume
/// the entire buffered window, but can't output anything until the size is >=
/// 0 (sizes are set to negative while they're pending calculation).
///
/// So SCAN takes input and buffers tokens and pending calculations, while
/// PRINT gobbles up completed calculations and tokens from the buffer. The
/// theory is that the two can never get more than 3N tokens apart, because
/// once there's "obviously" too much data to fit on a line, in a size
/// calculation, SCAN will write "infinity" to the size and let PRINT consume
/// it.
///
/// In this implementation (following the paper, again) the SCAN process is
/// the method called 'pretty_print', and the 'PRINT' process is the method
/// called 'print'.
pub struct Printer {
pub out: Box<io::Writer>,
buf_len: uint,
margin: int, // width of lines we're constrained to
space: int, // number of spaces left on line
left: uint, // index of left side of input stream
right: uint, // index of right side of input stream
token: Vec<Token> , // ring-buffr stream goes through
size: Vec<int> , // ring-buffer of calculated sizes
left_total: int, // running size of stream "...left"
right_total: int, // running size of stream "...right"
// pseudo-stack, really a ring too. Holds the
// primary-ring-buffers index of the Begin that started the
// current block, possibly with the most recent Break after that
// Begin (if there is any) on top of it. Stuff is flushed off the
// bottom as it becomes irrelevant due to the primary ring-buffer
// advancing.
/// Width of lines we're constrained to
margin: int,
/// Number of spaces left on line
space: int,
/// Index of left side of input stream
left: uint,
/// Index of right side of input stream
right: uint,
/// Ring-buffr stream goes through
token: Vec<Token> ,
/// Ring-buffer of calculated sizes
size: Vec<int> ,
/// Running size of stream "...left"
left_total: int,
/// Running size of stream "...right"
right_total: int,
/// Pseudo-stack, really a ring too. Holds the
/// primary-ring-buffers index of the Begin that started the
/// current block, possibly with the most recent Break after that
/// Begin (if there is any) on top of it. Stuff is flushed off the
/// bottom as it becomes irrelevant due to the primary ring-buffer
/// advancing.
scan_stack: Vec<uint> ,
scan_stack_empty: bool, // top==bottom disambiguator
top: uint, // index of top of scan_stack
bottom: uint, // index of bottom of scan_stack
// stack of blocks-in-progress being flushed by print
/// Top==bottom disambiguator
scan_stack_empty: bool,
/// Index of top of scan_stack
top: uint,
/// Index of bottom of scan_stack
bottom: uint,
/// Stack of blocks-in-progress being flushed by print
print_stack: Vec<PrintStackElem> ,
// buffered indentation to avoid writing trailing whitespace
/// Buffered indentation to avoid writing trailing whitespace
pending_indentation: int,
}

6
src/libsyntax/print/pprust.rs
View file

@ -88,9 +88,9 @@ pub static indent_unit: uint = 4u;
pub static default_columns: uint = 78u;
// Requires you to pass an input filename and reader so that
// it can scan the input text for comments and literals to
// copy forward.
/// Requires you to pass an input filename and reader so that
/// it can scan the input text for comments and literals to
/// copy forward.
pub fn print_crate<'a>(cm: &'a CodeMap,
span_diagnostic: &diagnostic::SpanHandler,
krate: &ast::Crate,

96
src/libsyntax/util/interner.rs
View file

@ -8,9 +8,9 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
// An "interner" is a data structure that associates values with uint tags and
// allows bidirectional lookup; i.e. given a value, one can easily find the
// type, and vice versa.
//! An "interner" is a data structure that associates values with uint tags and
//! allows bidirectional lookup; i.e. given a value, one can easily find the
//! type, and vice versa.
use ast::Name;
@ -52,7 +52,7 @@ impl<T: Eq + Hash + Clone + 'static> Interner<T> {
}
let mut vect = self.vect.borrow_mut();
let new_idx = (*vect).len() as Name;
let new_idx = Name((*vect).len() as u32);
(*map).insert(val.clone(), new_idx);
(*vect).push(val);
new_idx
@ -60,7 +60,7 @@ impl<T: Eq + Hash + Clone + 'static> Interner<T> {
pub fn gensym(&self, val: T) -> Name {
let mut vect = self.vect.borrow_mut();
let new_idx = (*vect).len() as Name;
let new_idx = Name((*vect).len() as u32);
// leave out of .map to avoid colliding
(*vect).push(val);
new_idx
@ -68,7 +68,7 @@ impl<T: Eq + Hash + Clone + 'static> Interner<T> {
pub fn get(&self, idx: Name) -> T {
let vect = self.vect.borrow();
(*(*vect).get(idx as uint)).clone()
(*(*vect).get(idx.uint())).clone()
}
pub fn len(&self) -> uint {
@ -155,7 +155,7 @@ impl StrInterner {
None => (),
}
let new_idx = self.len() as Name;
let new_idx = Name(self.len() as u32);
let val = RcStr::new(val);
map.insert(val.clone(), new_idx);
self.vect.borrow_mut().push(val);
@ -163,7 +163,7 @@ impl StrInterner {
}
pub fn gensym(&self, val: &str) -> Name {
let new_idx = self.len() as Name;
let new_idx = Name(self.len() as u32);
// leave out of .map to avoid colliding
self.vect.borrow_mut().push(RcStr::new(val));
new_idx
@ -180,23 +180,23 @@ impl StrInterner {
/// Create a gensym with the same name as an existing
/// entry.
pub fn gensym_copy(&self, idx : Name) -> Name {
let new_idx = self.len() as Name;
let new_idx = Name(self.len() as u32);
// leave out of map to avoid colliding
let mut vect = self.vect.borrow_mut();
let existing = (*vect.get(idx as uint)).clone();
let existing = (*vect.get(idx.uint())).clone();
vect.push(existing);
new_idx
}
pub fn get(&self, idx: Name) -> RcStr {
(*self.vect.borrow().get(idx as uint)).clone()
(*self.vect.borrow().get(idx.uint())).clone()
}
/// Returns this string with lifetime tied to the interner. Since
/// strings may never be removed from the interner, this is safe.
pub fn get_ref<'a>(&'a self, idx: Name) -> &'a str {
let vect = self.vect.borrow();
let s: &str = vect.get(idx as uint).as_slice();
let s: &str = vect.get(idx.uint()).as_slice();
unsafe {
mem::transmute(s)
}
@ -222,36 +222,38 @@ impl StrInterner {
#[cfg(test)]
mod tests {
use super::*;
use ast::Name;
#[test]
#[should_fail]
fn i1 () {
let i : Interner<RcStr> = Interner::new();
i.get(13);
i.get(Name(13));
}
#[test]
fn interner_tests () {
let i : Interner<RcStr> = Interner::new();
// first one is zero:
assert_eq!(i.intern(RcStr::new("dog")), 0);
assert_eq!(i.intern(RcStr::new("dog")), Name(0));
// re-use gets the same entry:
assert_eq!(i.intern(RcStr::new("dog")), 0);
assert_eq!(i.intern(RcStr::new("dog")), Name(0));
// different string gets a different #:
assert_eq!(i.intern(RcStr::new("cat")), 1);
assert_eq!(i.intern(RcStr::new("cat")), 1);
assert_eq!(i.intern(RcStr::new("cat")), Name(1));
assert_eq!(i.intern(RcStr::new("cat")), Name(1));
// dog is still at zero
assert_eq!(i.intern(RcStr::new("dog")), 0);
assert_eq!(i.intern(RcStr::new("dog")), Name(0));
// gensym gets 3
assert_eq!(i.gensym(RcStr::new("zebra") ), 2);
assert_eq!(i.gensym(RcStr::new("zebra") ), Name(2));
// gensym of same string gets new number :
assert_eq!(i.gensym (RcStr::new("zebra") ), 3);
assert_eq!(i.gensym (RcStr::new("zebra") ), Name(3));
// gensym of *existing* string gets new number:
assert_eq!(i.gensym(RcStr::new("dog")), 4);
assert_eq!(i.get(0), RcStr::new("dog"));
assert_eq!(i.get(1), RcStr::new("cat"));
assert_eq!(i.get(2), RcStr::new("zebra"));
assert_eq!(i.get(3), RcStr::new("zebra"));
assert_eq!(i.get(4), RcStr::new("dog"));
assert_eq!(i.gensym(RcStr::new("dog")), Name(4));
assert_eq!(i.get(Name(0)), RcStr::new("dog"));
assert_eq!(i.get(Name(1)), RcStr::new("cat"));
assert_eq!(i.get(Name(2)), RcStr::new("zebra"));
assert_eq!(i.get(Name(3)), RcStr::new("zebra"));
assert_eq!(i.get(Name(4)), RcStr::new("dog"));
}
#[test]
@ -261,39 +263,39 @@ mod tests {
RcStr::new("Bob"),
RcStr::new("Carol")
]);
assert_eq!(i.get(0), RcStr::new("Alan"));
assert_eq!(i.get(1), RcStr::new("Bob"));
assert_eq!(i.get(2), RcStr::new("Carol"));
assert_eq!(i.intern(RcStr::new("Bob")), 1);
assert_eq!(i.get(Name(0)), RcStr::new("Alan"));
assert_eq!(i.get(Name(1)), RcStr::new("Bob"));
assert_eq!(i.get(Name(2)), RcStr::new("Carol"));
assert_eq!(i.intern(RcStr::new("Bob")), Name(1));
}
#[test]
fn string_interner_tests() {
let i : StrInterner = StrInterner::new();
// first one is zero:
assert_eq!(i.intern("dog"), 0);
assert_eq!(i.intern("dog"), Name(0));
// re-use gets the same entry:
assert_eq!(i.intern ("dog"), 0);
assert_eq!(i.intern ("dog"), Name(0));
// different string gets a different #:
assert_eq!(i.intern("cat"), 1);
assert_eq!(i.intern("cat"), 1);
assert_eq!(i.intern("cat"), Name(1));
assert_eq!(i.intern("cat"), Name(1));
// dog is still at zero
assert_eq!(i.intern("dog"), 0);
assert_eq!(i.intern("dog"), Name(0));
// gensym gets 3
assert_eq!(i.gensym("zebra"), 2);
assert_eq!(i.gensym("zebra"), Name(2));
// gensym of same string gets new number :
assert_eq!(i.gensym("zebra"), 3);
assert_eq!(i.gensym("zebra"), Name(3));
// gensym of *existing* string gets new number:
assert_eq!(i.gensym("dog"), 4);
assert_eq!(i.gensym("dog"), Name(4));
// gensym tests again with gensym_copy:
assert_eq!(i.gensym_copy(2), 5);
assert_eq!(i.get(5), RcStr::new("zebra"));
assert_eq!(i.gensym_copy(2), 6);
assert_eq!(i.get(6), RcStr::new("zebra"));
assert_eq!(i.get(0), RcStr::new("dog"));
assert_eq!(i.get(1), RcStr::new("cat"));
assert_eq!(i.get(2), RcStr::new("zebra"));
assert_eq!(i.get(3), RcStr::new("zebra"));
assert_eq!(i.get(4), RcStr::new("dog"));
assert_eq!(i.gensym_copy(Name(2)), Name(5));
assert_eq!(i.get(Name(5)), RcStr::new("zebra"));
assert_eq!(i.gensym_copy(Name(2)), Name(6));
assert_eq!(i.get(Name(6)), RcStr::new("zebra"));
assert_eq!(i.get(Name(0)), RcStr::new("dog"));
assert_eq!(i.get(Name(1)), RcStr::new("cat"));
assert_eq!(i.get(Name(2)), RcStr::new("zebra"));
assert_eq!(i.get(Name(3)), RcStr::new("zebra"));
assert_eq!(i.get(Name(4)), RcStr::new("dog"));
}
}

38
src/libsyntax/util/parser_testing.rs
View file

@ -17,14 +17,14 @@ use parse::token;
use std::gc::Gc;
// map a string to tts, using a made-up filename:
/// Map a string to tts, using a made-up filename:
pub fn string_to_tts(source_str: String) -> Vec<ast::TokenTree> {
let ps = new_parse_sess();
filemap_to_tts(&ps,
string_to_filemap(&ps, source_str, "bogofile".to_string()))
}
// map string to parser (via tts)
/// Map string to parser (via tts)
pub fn string_to_parser<'a>(ps: &'a ParseSess, source_str: String) -> Parser<'a> {
new_parser_from_source_str(ps,
Vec::new(),
@ -40,51 +40,51 @@ fn with_error_checking_parse<T>(s: String, f: |&mut Parser| -> T) -> T {
x
}
// parse a string, return a crate.
/// Parse a string, return a crate.
pub fn string_to_crate (source_str : String) -> ast::Crate {
with_error_checking_parse(source_str, |p| {
p.parse_crate_mod()
})
}
// parse a string, return an expr
/// Parse a string, return an expr
pub fn string_to_expr (source_str : String) -> Gc<ast::Expr> {
with_error_checking_parse(source_str, |p| {
p.parse_expr()
})
}
// parse a string, return an item
/// Parse a string, return an item
pub fn string_to_item (source_str : String) -> Option<Gc<ast::Item>> {
with_error_checking_parse(source_str, |p| {
p.parse_item(Vec::new())
})
}
// parse a string, return a stmt
/// Parse a string, return a stmt
pub fn string_to_stmt(source_str : String) -> Gc<ast::Stmt> {
with_error_checking_parse(source_str, |p| {
p.parse_stmt(Vec::new())
})
}
// parse a string, return a pat. Uses "irrefutable"... which doesn't
// (currently) affect parsing.
/// Parse a string, return a pat. Uses "irrefutable"... which doesn't
/// (currently) affect parsing.
pub fn string_to_pat(source_str: String) -> Gc<ast::Pat> {
string_to_parser(&new_parse_sess(), source_str).parse_pat()
}
// convert a vector of strings to a vector of ast::Ident's
/// Convert a vector of strings to a vector of ast::Ident's
pub fn strs_to_idents(ids: Vec<&str> ) -> Vec<ast::Ident> {
ids.iter().map(|u| token::str_to_ident(*u)).collect()
}
// does the given string match the pattern? whitespace in the first string
// may be deleted or replaced with other whitespace to match the pattern.
// this function is unicode-ignorant; fortunately, the careful design of
// UTF-8 mitigates this ignorance. In particular, this function only collapses
// sequences of \n, \r, ' ', and \t, but it should otherwise tolerate unicode
// chars. Unsurprisingly, it doesn't do NKF-normalization(?).
/// Does the given string match the pattern? whitespace in the first string
/// may be deleted or replaced with other whitespace to match the pattern.
/// this function is unicode-ignorant; fortunately, the careful design of
/// UTF-8 mitigates this ignorance. In particular, this function only collapses
/// sequences of \n, \r, ' ', and \t, but it should otherwise tolerate unicode
/// chars. Unsurprisingly, it doesn't do NKF-normalization(?).
pub fn matches_codepattern(a : &str, b : &str) -> bool {
let mut idx_a = 0;
let mut idx_b = 0;
@ -122,9 +122,9 @@ pub fn matches_codepattern(a : &str, b : &str) -> bool {
}
}
// given a string and an index, return the first uint >= idx
// that is a non-ws-char or is outside of the legal range of
// the string.
/// Given a string and an index, return the first uint >= idx
/// that is a non-ws-char or is outside of the legal range of
/// the string.
fn scan_for_non_ws_or_end(a : &str, idx: uint) -> uint {
let mut i = idx;
let len = a.len();
@ -134,7 +134,7 @@ fn scan_for_non_ws_or_end(a : &str, idx: uint) -> uint {
i
}
// copied from lexer.
/// Copied from lexer.
pub fn is_whitespace(c: char) -> bool {
return c == ' ' || c == '\t' || c == '\r' || c == '\n';
}

32
src/libsyntax/visit.rs
View file

@ -8,6 +8,18 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
//! Context-passing AST walker. Each overridden visit method has full control
//! over what happens with its node, it can do its own traversal of the node's
//! children (potentially passing in different contexts to each), call
//! `visit::visit_*` to apply the default traversal algorithm (again, it can
//! override the context), or prevent deeper traversal by doing nothing.
//!
//! Note: it is an important invariant that the default visitor walks the body
//! of a function in "execution order" (more concretely, reverse post-order
//! with respect to the CFG implied by the AST), meaning that if AST node A may
//! execute before AST node B, then A is visited first. The borrow checker in
//! particular relies on this property.
//!
use abi::Abi;
use ast::*;
use ast;
@ -17,27 +29,15 @@ use owned_slice::OwnedSlice;
use std::gc::Gc;
// Context-passing AST walker. Each overridden visit method has full control
// over what happens with its node, it can do its own traversal of the node's
// children (potentially passing in different contexts to each), call
// visit::visit_* to apply the default traversal algorithm (again, it can
// override the context), or prevent deeper traversal by doing nothing.
//
// Note: it is an important invariant that the default visitor walks the body
// of a function in "execution order" (more concretely, reverse post-order
// with respect to the CFG implied by the AST), meaning that if AST node A may
// execute before AST node B, then A is visited first. The borrow checker in
// particular relies on this property.
pub enum FnKind<'a> {
// fn foo() or extern "Abi" fn foo()
/// fn foo() or extern "Abi" fn foo()
FkItemFn(Ident, &'a Generics, FnStyle, Abi),
// fn foo(&self)
/// fn foo(&self)
FkMethod(Ident, &'a Generics, &'a Method),
// |x, y| ...
// proc(x, y) ...
/// |x, y| ...
/// proc(x, y) ...
FkFnBlock,
}

16
src/test/compile-fail/lex-illegal-num-char-escape.rs → src/test/compile-fail/lex-bad-char-literals.rs
View file

@ -31,5 +31,19 @@ static s: &'static str =
static s2: &'static str =
"\u23q" //~ ERROR: illegal character in numeric character escape
//~^ ERROR: numeric character escape is too short
;
static c: char =
'\●' //~ ERROR: unknown character escape
;
static s: &'static str =
"\●" //~ ERROR: unknown character escape
;
// THIS MUST BE LAST, since unterminated character constants kill the lexer
static c: char =
'● //~ ERROR: unterminated character constant
;
//~^^ ERROR: numeric character escape is too short

13
src/test/compile-fail/lex-bad-fp-base-3.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let c = 0o3.0f32; //~ ERROR: octal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-4.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let d = 0o4e4; //~ ERROR: octal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-5.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let e = 0o5.0e5; //~ ERROR: octal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-6.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let f = 0o6e6f32; //~ ERROR: octal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-7.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let g = 0o7.0e7f64; //~ ERROR: octal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-8.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let h = 0x8.0e+9; //~ ERROR: hexadecimal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-base-9.rs
View file

@ -1,13 +0,0 @@
// Copyright 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.
fn main() {
let i = 0x9.0e-9; //~ ERROR: hexadecimal float literal is not supported
}

13
src/test/compile-fail/lex-bad-fp-lit.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static f: float =
1e+ //~ ERROR: scan_exponent: bad fp literal
;

35
src/test/compile-fail/lex-bad-numeric-literals.rs Normal file
View file

@ -0,0 +1,35 @@
// Copyright 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.
fn main() {
0o1.0; //~ ERROR: octal float literal is not supported
0o2f32; //~ ERROR: octal float literal is not supported
0o3.0f32; //~ ERROR: octal float literal is not supported
0o4e4; //~ ERROR: octal float literal is not supported
0o5.0e5; //~ ERROR: octal float literal is not supported
0o6e6f32; //~ ERROR: octal float literal is not supported
0o7.0e7f64; //~ ERROR: octal float literal is not supported
0x8.0e+9; //~ ERROR: hexadecimal float literal is not supported
0x9.0e-9; //~ ERROR: hexadecimal float literal is not supported
0o; //~ ERROR: no valid digits
1e+; //~ ERROR: expected at least one digit in exponent
0x539.0; //~ ERROR: hexadecimal float literal is not supported
99999999999999999999999999999999; //~ ERROR: int literal is too large
99999999999999999999999999999999u32; //~ ERROR: int literal is too large
0x; //~ ERROR: no valid digits
0xu32; //~ ERROR: no valid digits
0ou32; //~ ERROR: no valid digits
0bu32; //~ ERROR: no valid digits
0b; //~ ERROR: no valid digits
0o123f64; //~ ERROR: octal float literal is not supported
0o123.456; //~ ERROR: octal float literal is not supported
0b101f64; //~ ERROR: binary float literal is not supported
0b111.101; //~ ERROR: binary float literal is not supported
}

4
src/test/compile-fail/lex-bad-fp-base-1.rs → src/test/compile-fail/lex-bad-token.rs
View file

@ -8,6 +8,4 @@
// option. This file may not be copied, modified, or distributed
// except according to those terms.
fn main() {
let a = 0o1.0; //~ ERROR: octal float literal is not supported
}
● //~ ERROR: unknown start of token

13
src/test/compile-fail/lex-hex-float-lit.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static f: float =
0x539.0 //~ ERROR: hexadecimal float literal is not supported
;

13
src/test/compile-fail/lex-int-lit-too-large-2.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static i: int =
99999999999999999999999999999999u32 //~ ERROR: int literal is too large
;

13
src/test/compile-fail/lex-int-lit-too-large.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static i: int =
99999999999999999999999999999999 //~ ERROR: int literal is too large
;

13
src/test/compile-fail/lex-no-valid-digits-2.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static i: int =
0xu32 //~ ERROR: no valid digits
;

13
src/test/compile-fail/lex-no-valid-digits.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static i: int =
0x //~ ERROR: no valid digits
;

13
src/test/compile-fail/lex-unknown-char-escape.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static c: char =
'\●' //~ ERROR: unknown character escape
;

13
src/test/compile-fail/lex-unknown-start-tok.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
fn main() {
● //~ ERROR: unknown start of token
}

13
src/test/compile-fail/lex-unknown-str-escape.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static s: &'static str =
"\●" //~ ERROR: unknown character escape
;

13
src/test/compile-fail/lex-unterminated-char-const.rs
View file

@ -1,13 +0,0 @@
// Copyright 2013 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.
static c: char =
'● //~ ERROR: unterminated character constant
;

17
src/test/compile-fail/no-oct-float-literal.rs
View file

@ -1,17 +0,0 @@
// Copyright 2013 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.
// error-pattern:octal float literal is not supported
fn main() {
0o123f64;
0o123.456;
0o123p4f;
}

4
src/test/compile-fail/lex-bad-fp-base-2.rs → src/test/run-pass/string-escapes.rs
View file

@ -9,5 +9,7 @@
// except according to those terms.
fn main() {
let b = 0o2f32; //~ ERROR: octal float literal is not supported
let x = "\\\\\
";
assert!(x == r"\\"); // extraneous whitespace stripped
}