Add AbiSet and integrate it into the AST.

I believe this patch incorporates all expected syntax changes from extern
function reform (rust-lang#3678). You can now write things like:

    extern "<abi>" fn foo(s: S) -> T { ... }
    extern "<abi>" mod { ... }
    extern "<abi>" fn(S) -> T

The ABI for foreign functions is taken from this syntax (rather than from an
annotation).  We support the full ABI specification I described on the mailing
list.  The correct ABI is chosen based on the target architecture.

Calls by pointer to C functions are not yet supported, and the Rust type of
crust fns is still *u8.

Author: nikomatsakis

src/libcore/cast.rs

@@ -11,7 +11,7 @@
 pub mod rusti {
     #[abi = "rust-intrinsic"]
     #[link_name = "rusti"]
-    pub extern {
+    pub extern "rust-intrinsic" {
         fn forget<T>(+x: T);
         fn reinterpret_cast<T, U>(&&e: T) -> U;
     }

src/libcore/libc.rs

@@ -1617,7 +1617,7 @@ pub mod funcs {
             use libc::types::os::arch::extra::{HANDLE};
 
             #[abi = "stdcall"]
-            pub extern {
+            pub extern "stdcall" {
                 unsafe fn GetEnvironmentVariableW(n: LPCWSTR,
                                                   v: LPWSTR,
                                                   nsize: DWORD)

src/libcore/os.rs

@@ -944,7 +944,7 @@ pub fn errno() -> uint {
 
     #[link_name = "kernel32"]
     #[abi = "stdcall"]
-    extern {
+    extern "stdcall" {
         unsafe fn GetLastError() -> DWORD;
     }
 
@@ -1006,7 +1006,7 @@ pub fn last_os_error() -> ~str {
 
         #[link_name = "kernel32"]
         #[abi = "stdcall"]
-        extern {
+        extern "stdcall" {
             unsafe fn FormatMessageA(flags: DWORD, lpSrc: LPVOID,
                                      msgId: DWORD, langId: DWORD,
                                      buf: LPSTR, nsize: DWORD,

src/libcore/ptr.rs

@@ -44,7 +44,7 @@ pub mod libc_ {
 
 pub mod rusti {
     #[abi = "rust-intrinsic"]
-    pub extern {
+    pub extern "rust-intrinsic" {
         fn addr_of<T>(&&val: T) -> *T;
     }
 }

src/libcore/rt/thread_local_storage.rs

@@ -73,7 +73,7 @@ pub unsafe fn get(key: Key) -> *mut c_void {
 
 #[cfg(windows)]
 #[abi = "stdcall"]
-extern {
+extern "stdcall" {
        fn TlsAlloc() -> DWORD;
        fn TlsSetValue(dwTlsIndex: DWORD, lpTlsvalue: LPVOID) -> BOOL;
        fn TlsGetValue(dwTlsIndex: DWORD) -> LPVOID;

src/libcore/stackwalk.rs

@@ -94,7 +94,7 @@ pub mod rustrt {
 
 pub mod rusti {
     #[abi = "rust-intrinsic"]
-    pub extern {
+    pub extern "rust-intrinsic" {
         pub fn frame_address(f: &once fn(++x: *u8));
     }
 }

src/libcore/sys.rs

@@ -39,7 +39,7 @@ pub struct Closure {
 
 pub mod rusti {
     #[abi = "rust-intrinsic"]
-    pub extern {
+    pub extern "rust-intrinsic" {
         fn get_tydesc<T>() -> *();
         fn size_of<T>() -> uint;
         fn pref_align_of<T>() -> uint;

src/libcore/unstable/intrinsics.rs

@@ -15,7 +15,7 @@ The intrinsics are defined in librustc/middle/trans/foreign.rs.
 */
 
 #[abi = "rust-intrinsic"]
-pub extern {
+pub extern "rust-intrinsic" {
     pub fn atomic_cxchg(dst: &mut int, old: int, src: int) -> int;
     pub fn atomic_cxchg_acq(dst: &mut int, old: int, src: int) -> int;
     pub fn atomic_cxchg_rel(dst: &mut int, old: int, src: int) -> int;

src/librustc/back/link.rs

@@ -97,7 +97,7 @@ pub mod jit {
     pub mod rusti {
         #[nolink]
         #[abi = "rust-intrinsic"]
-        pub extern {
+        pub extern "rust-intrinsic" {
             pub fn morestack_addr() -> *();
         }
     }

src/librustc/driver/driver.rs

@@ -34,6 +34,7 @@ use std::getopts::groups::{optopt, optmulti, optflag, optflagopt, getopts};
 use std::getopts::{opt_present};
 use std::getopts;
 use syntax::ast;
+use syntax::abi;
 use syntax::attr;
 use syntax::codemap;
 use syntax::diagnostic;
@@ -86,10 +87,10 @@ pub fn default_configuration(sess: Session, +argv0: ~str, input: input) ->
     // ARM is bi-endian, however using NDK seems to default
     // to little-endian unless a flag is provided.
     let (end,arch,wordsz) = match sess.targ_cfg.arch {
-      session::arch_x86 => (~"little",~"x86",~"32"),
-      session::arch_x86_64 => (~"little",~"x86_64",~"64"),
-      session::arch_arm => (~"little",~"arm",~"32"),
-      session::arch_mips => (~"little",~"arm",~"32")
+        abi::X86 => (~"little",~"x86",~"32"),
+        abi::X86_64 => (~"little",~"x86_64",~"64"),
+        abi::Arm => (~"little",~"arm",~"32"),
+        abi::Mips => (~"little",~"arm",~"32")
     };
 
     return ~[ // Target bindings.
@@ -309,7 +310,7 @@ pub fn compile_rest(sess: Session, cfg: ast::crate_cfg,
     };
 
     // NOTE: Android hack
-    if sess.targ_cfg.arch == session::arch_arm &&
+    if sess.targ_cfg.arch == abi::Arm &&
             (sess.opts.output_type == link::output_type_object ||
              sess.opts.output_type == link::output_type_exe) {
         let output_type = link::output_type_assembly;
@@ -454,20 +455,20 @@ pub fn get_os(triple: &str) -> Option<session::os> {
         } else { None }
 }
 
-pub fn get_arch(triple: &str) -> Option<session::arch> {
+pub fn get_arch(triple: &str) -> Option<abi::Architecture> {
     if str::contains(triple, ~"i386") ||
         str::contains(triple, ~"i486") ||
                str::contains(triple, ~"i586") ||
                str::contains(triple, ~"i686") ||
                str::contains(triple, ~"i786") {
-            Some(session::arch_x86)
+            Some(abi::X86)
         } else if str::contains(triple, ~"x86_64") {
-            Some(session::arch_x86_64)
+            Some(abi::X86_64)
         } else if str::contains(triple, ~"arm") ||
                       str::contains(triple, ~"xscale") {
-            Some(session::arch_arm)
+            Some(abi::Arm)
         } else if str::contains(triple, ~"mips") {
-            Some(session::arch_mips)
+            Some(abi::Mips)
         } else { None }
 }
 
@@ -484,16 +485,16 @@ pub fn build_target_config(sopts: @session::options,
                           ~"unknown architecture: " + sopts.target_triple)
     };
     let (int_type, uint_type, float_type) = match arch {
-      session::arch_x86 => (ast::ty_i32, ast::ty_u32, ast::ty_f64),
-      session::arch_x86_64 => (ast::ty_i64, ast::ty_u64, ast::ty_f64),
-      session::arch_arm => (ast::ty_i32, ast::ty_u32, ast::ty_f64),
-      session::arch_mips => (ast::ty_i32, ast::ty_u32, ast::ty_f64)
+      abi::X86 => (ast::ty_i32, ast::ty_u32, ast::ty_f64),
+      abi::X86_64 => (ast::ty_i64, ast::ty_u64, ast::ty_f64),
+      abi::Arm => (ast::ty_i32, ast::ty_u32, ast::ty_f64),
+      abi::Mips => (ast::ty_i32, ast::ty_u32, ast::ty_f64)
     };
     let target_strs = match arch {
-      session::arch_x86 => x86::get_target_strs(os),
-      session::arch_x86_64 => x86_64::get_target_strs(os),
-      session::arch_arm => arm::get_target_strs(os),
-      session::arch_mips => mips::get_target_strs(os)
+      abi::X86 => x86::get_target_strs(os),
+      abi::X86_64 => x86_64::get_target_strs(os),
+      abi::Arm => arm::get_target_strs(os),
+      abi::Mips => mips::get_target_strs(os)
     };
     let target_cfg = @session::config {
         os: os,

src/librustc/driver/session.rs

@@ -25,19 +25,17 @@ use syntax::codemap::span;
 use syntax::diagnostic;
 use syntax::parse::ParseSess;
 use syntax::{ast, codemap};
+use syntax::abi;
 use syntax;
 
 #[deriving(Eq)]
 pub enum os { os_win32, os_macos, os_linux, os_android, os_freebsd, }
 
-#[deriving(Eq)]
-pub enum arch { arch_x86, arch_x86_64, arch_arm, arch_mips, }
-
 pub enum crate_type { bin_crate, lib_crate, unknown_crate, }
 
 pub struct config {
     os: os,
-    arch: arch,
+    arch: abi::Architecture,
     target_strs: target_strs::t,
     int_type: int_ty,
     uint_type: uint_ty,

src/librustc/front/config.rs

@@ -92,7 +92,7 @@ fn fold_foreign_mod(
         nm.view_items.filter_mapped(|a| filter_view_item(cx, *a));
     ast::foreign_mod {
         sort: nm.sort,
-        abi: nm.abi,
+        abis: nm.abis,
         view_items: vec::map(filtered_view_items, |x| fld.fold_view_item(*x)),
         items: filtered_items
     }

src/librustc/front/intrinsic.rs

@@ -126,7 +126,7 @@ pub mod intrinsic {
         use super::{TyDesc, TyVisitor};
 
         #[abi = "rust-intrinsic"]
-        pub extern {
+        pub extern "rust-intrinsic" {
             pub fn get_tydesc<T>() -> *();
             pub fn visit_tydesc(++td: *TyDesc, &&tv: @TyVisitor);
         }

src/librustc/front/test.rs

@@ -146,7 +146,7 @@ fn fold_item(cx: @mut TestCtxt, &&i: @ast::item, fld: @fold::ast_fold)
 
     if is_test_fn(i) || is_bench_fn(i) {
         match i.node {
-          ast::item_fn(_, purity, _, _) if purity == ast::unsafe_fn => {
+          ast::item_fn(_, purity, _, _, _) if purity == ast::unsafe_fn => {
             let sess = cx.sess;
             sess.span_fatal(
                 i.span,
@@ -178,7 +178,7 @@ fn is_test_fn(i: @ast::item) -> bool {
 
     fn has_test_signature(i: @ast::item) -> bool {
         match &i.node {
-          &ast::item_fn(ref decl, _, ref generics, _) => {
+          &ast::item_fn(ref decl, _, _, ref generics, _) => {
             let no_output = match decl.output.node {
                 ast::ty_nil => true,
                 _ => false
@@ -200,7 +200,7 @@ fn is_bench_fn(i: @ast::item) -> bool {
 
     fn has_test_signature(i: @ast::item) -> bool {
         match i.node {
-            ast::item_fn(ref decl, _, ref generics, _) => {
+            ast::item_fn(ref decl, _, _, ref generics, _) => {
                 let input_cnt = vec::len(decl.inputs);
                 let no_output = match decl.output.node {
                     ast::ty_nil => true,

src/librustc/metadata/creader.rs

@@ -156,12 +156,8 @@ fn visit_view_item(e: @mut Env, i: @ast::view_item) {
 fn visit_item(e: @mut Env, i: @ast::item) {
     match i.node {
       ast::item_foreign_mod(ref fm) => {
-        match attr::foreign_abi(i.attrs) {
-          either::Right(abi) => {
-            if abi != ast::foreign_abi_cdecl &&
-               abi != ast::foreign_abi_stdcall { return; }
-          }
-          either::Left(ref msg) => e.diag.span_fatal(i.span, (*msg))
+        if fm.abis.is_rust() || fm.abis.is_intrinsic() {
+            return;
         }
 
         let cstore = e.cstore;

src/librustc/metadata/encoder.rs

@@ -35,6 +35,7 @@ use std::oldmap::HashMap;
 use std::serialize::Encodable;
 use std::{ebml, oldmap};
 use std;
+use syntax::abi::AbiSet;
 use syntax::ast::*;
 use syntax::ast;
 use syntax::ast_map;
@@ -657,7 +658,7 @@ fn encode_info_for_item(ecx: @EncodeContext, ebml_w: writer::Encoder,
         (ecx.encode_inlined_item)(ecx, ebml_w, path, ii_item(item));
         ebml_w.end_tag();
       }
-      item_fn(_, purity, ref generics, _) => {
+      item_fn(_, purity, _, ref generics, _) => {
         add_to_index();
         ebml_w.start_tag(tag_items_data_item);
         encode_def_id(ebml_w, local_def(item.id));
@@ -983,7 +984,7 @@ fn encode_info_for_foreign_item(ecx: @EncodeContext,
                                 nitem: @foreign_item,
                                 index: @mut ~[entry<int>],
                                 +path: ast_map::path,
-                                abi: foreign_abi) {
+                                abi: AbiSet) {
     if !reachable(ecx, nitem.id) { return; }
     index.push(entry { val: nitem.id, pos: ebml_w.writer.tell() });
 
@@ -994,7 +995,7 @@ fn encode_info_for_foreign_item(ecx: @EncodeContext,
         encode_family(ebml_w, purity_fn_family(purity));
         encode_type_param_bounds(ebml_w, ecx, &generics.ty_params);
         encode_type(ecx, ebml_w, node_id_to_type(ecx.tcx, nitem.id));
-        if abi == foreign_abi_rust_intrinsic {
+        if abi.is_intrinsic() {
             (ecx.encode_inlined_item)(ecx, ebml_w, path, ii_foreign(nitem));
         } else {
             encode_symbol(ecx, ebml_w, nitem.id);

src/librustc/metadata/tydecode.rs

@@ -21,6 +21,8 @@ use middle::ty;
 use core::str;
 use core::uint;
 use core::vec;
+use syntax::abi::AbiSet;
+use syntax::abi;
 use syntax::ast;
 use syntax::ast::*;
 use syntax::codemap::{respan, dummy_sp};
@@ -74,17 +76,28 @@ fn next_byte(st: @mut PState) -> u8 {
     return b;
 }
 
+fn scan<R>(st: &mut PState, is_last: &fn(char) -> bool,
+           op: &fn(&[u8]) -> R) -> R
+{
+    let start_pos = st.pos;
+    debug!("scan: '%c' (start)", st.data[st.pos] as char);
+    while !is_last(st.data[st.pos] as char) {
+        st.pos += 1;
+        debug!("scan: '%c'", st.data[st.pos] as char);
+    }
+    let end_pos = st.pos;
+    st.pos += 1;
+    return op(st.data.slice(start_pos, end_pos));
+}
+
 pub fn parse_ident(st: @mut PState, last: char) -> ast::ident {
     fn is_last(b: char, c: char) -> bool { return c == b; }
     return parse_ident_(st, |a| is_last(last, a) );
 }
 
 fn parse_ident_(st: @mut PState, is_last: @fn(char) -> bool) ->
    ast::ident {
-    let mut rslt = ~"";
-    while !is_last(peek(st)) {
-        rslt += str::from_byte(next_byte(st));
-    }
+    let rslt = scan(st, is_last, str::from_bytes);
     return st.tcx.sess.ident_of(rslt);
 }
 
@@ -415,11 +428,17 @@ fn parse_purity(c: char) -> purity {
     }
 }
 
-fn parse_abi(c: char) -> Abi {
-    match c {
-      'r' => ast::RustAbi,
-      _ => fail!(fmt!("parse_abi: bad ABI '%c'", c))
+fn parse_abi_set(st: @mut PState) -> AbiSet {
+    fail_unless!(next(st) == '[');
+    let mut abis = AbiSet::empty();
+    while peek(st) != ']' {
+        // FIXME(#5422) str API should not force this copy
+        let abi_str = scan(st, |c| c == ',', str::from_bytes);
+        let abi = abi::lookup(abi_str).expect(abi_str);
+        abis.add(abi);
     }
+    fail_unless!(next(st) == ']');
+    return abis;
 }
 
 fn parse_onceness(c: char) -> ast::Onceness {
@@ -460,11 +479,11 @@ fn parse_closure_ty(st: @mut PState, conv: conv_did) -> ty::ClosureTy {
 
 fn parse_bare_fn_ty(st: @mut PState, conv: conv_did) -> ty::BareFnTy {
     let purity = parse_purity(next(st));
-    let abi = parse_abi(next(st));
+    let abi = parse_abi_set(st);
     let sig = parse_sig(st, conv);
     ty::BareFnTy {
         purity: purity,
-        abi: abi,
+        abis: abi,
         sig: sig
     }
 }