---

yaml
---
r: 153851
b: refs/heads/try2
c: b2bd998
h: refs/heads/master
i:
  153849: 8fc4aa0
  153847: 632c11a
v: v3

Author: bors

[refs]

@@ -5,7 +5,7 @@ refs/heads/snap-stage3: 78a7676898d9f80ab540c6df5d4c9ce35bb50463
 refs/heads/try: 519addf6277dbafccbb4159db4b710c37eaa2ec5
 refs/tags/release-0.1: 1f5c5126e96c79d22cb7862f75304136e204f105
 refs/heads/ndm: f3868061cd7988080c30d6d5bf352a5a5fe2460b
-refs/heads/try2: 3f568464608ad39f1d620b81d97f844015f971a1
+refs/heads/try2: b2bd9986072a010cd68a7f0a814ad99038050983
 refs/heads/dist-snap: ba4081a5a8573875fed17545846f6f6902c8ba8d
 refs/tags/release-0.2: c870d2dffb391e14efb05aa27898f1f6333a9596
 refs/tags/release-0.3: b5f0d0f648d9a6153664837026ba1be43d3e2503

branches/try2/src/doc/guide-container.md

@@ -118,7 +118,7 @@ differently.
 ## Container iterators
 
 Containers implement iteration over the contained elements by returning an
-iterator object. For example, vector slices several iterators available:
+iterator object. For example, for vector slices several iterators are available:
 
 * `iter()` for immutable references to the elements
 * `mut_iter()` for mutable references to the elements

branches/try2/src/grammar/raw-string-literal-ambiguity.md

@@ -1,7 +1,11 @@
 Rust's lexical grammar is not context-free. Raw string literals are the source
 of the problem. Informally, a raw string literal is an `r`, followed by `N`
 hashes (where N can be zero), a quote, any characters, then a quote followed
-by `N` hashes. This grammar describes this as best possible:
+by `N` hashes. Critically, once inside the first pair of quotes,
+another quote cannot be followed by `N` consecutive hashes. e.g.
+`r###""###"###` is invalid.
+
+This grammar describes this as best possible:
 
     R -> 'r' S
     S -> '"' B '"'
@@ -22,8 +26,39 @@ accepted as one by the above grammar, using the derivation:
 (Where `T : U` means the rule `T` is applied, and `U` is the remainder of the
 string.) The difficulty arises from the fact that it is fundamentally
 context-sensitive. In particular, the context needed is the number of hashes.
-I know of no way to resolve this, but also have not come up with a proof that
-it is not context sensitive. Such a proof would probably use the pumping lemma
-for context-free languages, but I (cmr) could not come up with a proof after
-spending a few hours on it, and decided my time best spent elsewhere. Pull
-request welcome!
+
+To prove that Rust's string literals are not context-free, we will use
+the fact that context-free languages are closed under intersection with
+regular languages, and the
+[pumping lemma for context-free languages](https://en.wikipedia.org/wiki/Pumping_lemma_for_context-free_languages).
+
+Consider the regular language `R = r#+""#*"#+`. If Rust's raw string literals are
+context-free, then their intersection with `R`, `R'`, should also be context-free.
+Therefore, to prove that raw string literals are not context-free,
+it is sufficient to prove that `R'` is not context-free.
+
+The language `R'` is `{r#^n""#^m"#^n | m < n}`.
+
+Assume `R'` *is* context-free. Then `R'` has some pumping length `p > 0` for which
+the pumping lemma applies. Consider the following string `s` in `R'`:
+
+`r#^p""#^{p-1}"#^p`
+
+e.g. for `p = 2`: `s = r##""#"##`
+
+Then `s = uvwxy` for some choice of `uvwxy` such that `vx` is non-empty,
+`|vwx| < p+1`, and `uv^iwx^iy` is in `R'` for all `i >= 0`.
+
+Neither `v` nor `x` can contain a `"` or `r`, as the number of these characters
+in any string in `R'` is fixed. So `v` and `x` contain only hashes.
+Consequently, of the three sequences of hashes, `v` and `x` combined
+can only pump two of them.
+If we ever choose the central sequence of hashes, then one of the outer sequences
+will not grow when we pump, leading to an imbalance between the outer sequences.
+Therefore, we must pump both outer sequences of hashes. However,
+there are `p+2` characters between these two sequences of hashes, and `|vwx|` must
+be less than `p+1`. Therefore we have a contradiction, and `R'` must not be
+context-free.
+
+Since `R'` is not context-free, it follows that the Rust's raw string literals
+must not be context-free.

branches/try2/src/libcollections/ringbuf.rs

@@ -19,6 +19,8 @@ use core::cmp;
 use core::default::Default;
 use core::fmt;
 use core::iter::RandomAccessIterator;
+use core::iter;
+use std::hash::{Writer, Hash};
 
 use {Deque, Collection, Mutable, MutableSeq};
 use vec::Vec;
@@ -450,6 +452,21 @@ impl<A: PartialEq> PartialEq for RingBuf<A> {
     }
 }
 
+impl<A: PartialOrd> PartialOrd for RingBuf<A> {
+    fn partial_cmp(&self, other: &RingBuf<A>) -> Option<Ordering> {
+        iter::order::partial_cmp(self.iter(), other.iter())
+    }
+}
+
+impl<S: Writer, A: Hash<S>> Hash<S> for RingBuf<A> {
+    fn hash(&self, state: &mut S) {
+        self.len().hash(state);
+        for elt in self.iter() {
+            elt.hash(state);
+        }
+    }
+}
+
 impl<A> FromIterator<A> for RingBuf<A> {
     fn from_iter<T: Iterator<A>>(iterator: T) -> RingBuf<A> {
         let (lower, _) = iterator.size_hint();
@@ -485,6 +502,7 @@ mod tests {
     use std::fmt::Show;
     use std::prelude::*;
     use std::gc::{GC, Gc};
+    use std::hash;
     use test::Bencher;
     use test;
 
@@ -912,6 +930,37 @@ mod tests {
         assert!(e == RingBuf::new());
     }
 
+    #[test]
+    fn test_hash() {
+      let mut x = RingBuf::new();
+      let mut y = RingBuf::new();
+
+      x.push(1i);
+      x.push(2);
+      x.push(3);
+
+      y.push(0i);
+      y.push(1i);
+      y.pop_front();
+      y.push(2);
+      y.push(3);
+
+      assert!(hash::hash(&x) == hash::hash(&y));
+    }
+
+    #[test]
+    fn test_ord() {
+        let x = RingBuf::new();
+        let mut y = RingBuf::new();
+        y.push(1i);
+        y.push(2);
+        y.push(3);
+        assert!(x < y);
+        assert!(y > x);
+        assert!(x <= x);
+        assert!(x >= x);
+    }
+
     #[test]
     fn test_show() {
         let ringbuf: RingBuf<int> = range(0i, 10).collect();

branches/try2/src/libcore/failure.rs

@@ -37,7 +37,7 @@ use intrinsics;
 #[lang="fail_"]
 fn fail_(expr: &'static str, file: &'static str, line: uint) -> ! {
     format_args!(|args| -> () {
-        begin_unwind(args, file, line);
+        begin_unwind(args, &(file, line));
     }, "{}", expr);
 
     unsafe { intrinsics::abort() }
@@ -48,18 +48,20 @@ fn fail_(expr: &'static str, file: &'static str, line: uint) -> ! {
 fn fail_bounds_check(file: &'static str, line: uint,
                      index: uint, len: uint) -> ! {
     format_args!(|args| -> () {
-        begin_unwind(args, file, line);
+        begin_unwind(args, &(file, line));
     }, "index out of bounds: the len is {} but the index is {}", len, index);
     unsafe { intrinsics::abort() }
 }
 
 #[cold]
-pub fn begin_unwind(fmt: &fmt::Arguments, file: &'static str, line: uint) -> ! {
+pub fn begin_unwind(fmt: &fmt::Arguments, file_line: &(&'static str, uint)) -> ! {
     #[allow(ctypes)]
     extern {
         #[lang = "begin_unwind"]
         fn begin_unwind(fmt: &fmt::Arguments, file: &'static str,
                         line: uint) -> !;
     }
+    let (file, line) = *file_line;
     unsafe { begin_unwind(fmt, file, line) }
 }
+

branches/try2/src/libcore/macros.rs

@@ -33,7 +33,8 @@ macro_rules! fail(
         // up with the number of calls to fail!()
         #[inline(always)]
         fn run_fmt(fmt: &::std::fmt::Arguments) -> ! {
-            ::core::failure::begin_unwind(fmt, file!(), line!())
+            static FILE_LINE: (&'static str, uint) = (file!(), line!());
+            ::core::failure::begin_unwind(fmt, &FILE_LINE)
         }
         format_args!(run_fmt, $fmt, $($arg)*)
     });

branches/try2/src/librustc/metadata/common.rs

@@ -227,3 +227,9 @@ pub static tag_region_param_def_index: uint = 0x94;
 pub static tag_unboxed_closures: uint = 0x95;
 pub static tag_unboxed_closure: uint = 0x96;
 pub static tag_unboxed_closure_type: uint = 0x97;
+
+pub static tag_struct_fields: uint = 0x98;
+pub static tag_struct_field: uint = 0x99;
+pub static tag_struct_field_id: uint = 0x9a;
+
+pub static tag_attribute_is_sugared_doc: uint = 0x9b;

branches/try2/src/librustc/metadata/csearch.rs

@@ -29,6 +29,8 @@ use syntax::attr;
 use syntax::diagnostic::expect;
 use syntax::parse::token;
 
+use std::collections::hashmap::HashMap;
+
 pub struct StaticMethodInfo {
     pub ident: ast::Ident,
     pub def_id: ast::DefId,
@@ -192,6 +194,12 @@ pub fn get_struct_fields(cstore: &cstore::CStore,
     decoder::get_struct_fields(cstore.intr.clone(), &*cdata, def.node)
 }
 
+pub fn get_struct_field_attrs(cstore: &cstore::CStore, def: ast::DefId) -> HashMap<ast::NodeId,
+        Vec<ast::Attribute>> {
+    let cdata = cstore.get_crate_data(def.krate);
+    decoder::get_struct_field_attrs(&*cdata)
+}
+
 pub fn get_type(tcx: &ty::ctxt,
                 def: ast::DefId)
              -> ty::Polytype {

branches/try2/src/librustc/metadata/decoder.rs

@@ -34,6 +34,7 @@ use std::hash::Hash;
 use std::hash;
 use std::io::extensions::u64_from_be_bytes;
 use std::io;
+use std::collections::hashmap::HashMap;
 use std::rc::Rc;
 use std::u64;
 use serialize::ebml::reader;
@@ -963,6 +964,19 @@ pub fn get_item_attrs(cdata: Cmd,
     f(get_attributes(item));
 }
 
+pub fn get_struct_field_attrs(cdata: Cmd) -> HashMap<ast::NodeId, Vec<ast::Attribute>> {
+    let data = ebml::Doc::new(cdata.data());
+    let fields = reader::get_doc(data, tag_struct_fields);
+    let mut map = HashMap::new();
+    reader::tagged_docs(fields, tag_struct_field, |field| {
+        let id = reader::doc_as_u32(reader::get_doc(field, tag_struct_field_id));
+        let attrs = get_attributes(field);
+        map.insert(id, attrs);
+        true
+    });
+    map
+}
+
 fn struct_field_family_to_visibility(family: Family) -> ast::Visibility {
     match family {
       PublicField => ast::Public,
@@ -1042,6 +1056,9 @@ fn get_attributes(md: ebml::Doc) -> Vec<ast::Attribute> {
     match reader::maybe_get_doc(md, tag_attributes) {
       Some(attrs_d) => {
         reader::tagged_docs(attrs_d, tag_attribute, |attr_doc| {
+            let is_sugared_doc = reader::doc_as_u8(
+                reader::get_doc(attr_doc, tag_attribute_is_sugared_doc)
+            ) == 1;
             let meta_items = get_meta_items(attr_doc);
             // Currently it's only possible to have a single meta item on
             // an attribute
@@ -1053,7 +1070,7 @@ fn get_attributes(md: ebml::Doc) -> Vec<ast::Attribute> {
                         id: attr::mk_attr_id(),
                         style: ast::AttrOuter,
                         value: meta_item,
-                        is_sugared_doc: false,
+                        is_sugared_doc: is_sugared_doc,
                     },
                     span: codemap::DUMMY_SP
                 });

branches/try2/src/librustc/metadata/encoder.rs

@@ -1490,6 +1490,7 @@ fn encode_attributes(ebml_w: &mut Encoder, attrs: &[Attribute]) {
     ebml_w.start_tag(tag_attributes);
     for attr in attrs.iter() {
         ebml_w.start_tag(tag_attribute);
+        ebml_w.wr_tagged_u8(tag_attribute_is_sugared_doc, attr.node.is_sugared_doc as u8);
         encode_meta_item(ebml_w, attr.node.value);
         ebml_w.end_tag();
     }
@@ -1644,6 +1645,29 @@ fn encode_unboxed_closures<'a>(
     ebml_w.end_tag();
 }
 
+fn encode_struct_field_attrs(ebml_w: &mut Encoder, krate: &Crate) {
+    struct StructFieldVisitor<'a, 'b> {
+        ebml_w: &'a mut Encoder<'b>,
+    }
+
+    impl<'a, 'b> Visitor<()> for StructFieldVisitor<'a, 'b> {
+        fn visit_struct_field(&mut self, field: &ast::StructField, _: ()) {
+            self.ebml_w.start_tag(tag_struct_field);
+            self.ebml_w.wr_tagged_u32(tag_struct_field_id, field.node.id);
+            encode_attributes(self.ebml_w, field.node.attrs.as_slice());
+            self.ebml_w.end_tag();
+        }
+    }
+
+    ebml_w.start_tag(tag_struct_fields);
+    visit::walk_crate(&mut StructFieldVisitor {
+        ebml_w: ebml_w
+    }, krate, ());
+    ebml_w.end_tag();
+}
+
+
+
 struct ImplVisitor<'a,'b,'c> {
     ecx: &'a EncodeContext<'b>,
     ebml_w: &'a mut Encoder<'c>,
@@ -1928,6 +1952,8 @@ fn encode_metadata_inner(wr: &mut MemWriter, parms: EncodeParams, krate: &Crate)
     stats.index_bytes = ebml_w.writer.tell().unwrap() - i;
     ebml_w.end_tag();
 
+    encode_struct_field_attrs(&mut ebml_w, krate);
+
     stats.total_bytes = ebml_w.writer.tell().unwrap();
 
     if tcx.sess.meta_stats() {

branches/try2/src/librustc/middle/ty.rs

@@ -4071,7 +4071,7 @@ pub fn lookup_struct_fields(cx: &ctxt, did: ast::DefId) -> Vec<field_ty> {
 
         let len = results.as_slice().iter().map(|x| x.len()).sum();
         let mut result: Vec<field_ty> = Vec::with_capacity(len);
-        result.extend(results.as_slice().iter().flat_map(|rs| rs.iter().map(|&f| f)));
+        result.extend(results.as_slice().iter().flat_map(|rs| rs.iter().map(|f| f.clone())));
         assert!(result.len() == len);
         result
     } else {
@@ -4085,7 +4085,7 @@ pub fn lookup_struct_field(cx: &ctxt,
                         -> field_ty {
     let r = lookup_struct_fields(cx, parent);
     match r.iter().find(|f| f.id.node == field_id.node) {
-        Some(t) => *t,
+        Some(t) => t.clone(),
         None => cx.sess.bug("struct ID not found in parent's fields")
     }
 }

branches/try2/src/librustc/middle/typeck/check/_match.rs

@@ -322,7 +322,7 @@ pub fn check_struct_pat_fields(pcx: &pat_ctxt,
             }
             Some(&(index, ref mut used)) => {
                 *used = true;
-                let class_field = *class_fields.get(index);
+                let class_field = class_fields.get(index).clone();
                 let field_type = ty::lookup_field_type(tcx,
                                                        class_id,
                                                        class_field.id,

branches/try2/src/librustc/middle/typeck/coherence.rs

@@ -60,29 +60,25 @@ fn get_base_type(inference_context: &InferCtxt,
                  span: Span,
                  original_type: t)
                  -> Option<t> {
-    let resolved_type;
-    match resolve_type(inference_context,
-                       Some(span),
-                       original_type,
-                       resolve_ivar) {
-        Ok(resulting_type) if !type_is_ty_var(resulting_type) => {
-            resolved_type = resulting_type;
-        }
+    let resolved_type = match resolve_type(inference_context,
+                                           Some(span),
+                                           original_type,
+                                           resolve_ivar) {
+        Ok(resulting_type) if !type_is_ty_var(resulting_type) => resulting_type,
         _ => {
             inference_context.tcx.sess.span_fatal(span,
                                                   "the type of this value must be known in order \
                                                    to determine the base type");
         }
-    }
+    };
 
     match get(resolved_type).sty {
         ty_enum(..) | ty_struct(..) | ty_unboxed_closure(..) => {
             debug!("(getting base type) found base type");
             Some(resolved_type)
         }
-        // FIXME(14865) I would prefere to use `_` here, but that causes a
-        // compiler error.
-        ty_uniq(_) | ty_rptr(_, _) | ty_trait(..) if ty::type_is_trait(resolved_type) => {
+
+        _ if ty::type_is_trait(resolved_type) => {
             debug!("(getting base type) found base type (trait)");
             Some(resolved_type)
         }

branches/try2/src/librustdoc/clean/inline.rs

@@ -209,7 +209,7 @@ fn build_struct(tcx: &ty::ctxt, did: ast::DefId) -> clean::Struct {
             _ => doctree::Plain,
         },
         generics: (&t.generics, subst::TypeSpace).clean(),
-        fields: fields.iter().map(|f| f.clean()).collect(),
+        fields: fields.clean(),
         fields_stripped: false,
     }
 }