auto merge of #15085 : brson/rust/stridx, r=alexcrichton

Being able to index into the bytes of a string encourages
poor UTF-8 hygiene. To get a view of `&[u8]` from either
a `String` or `&str` slice, use the `as_bytes()` method.

Closes #12710.

[breaking-change]

If the diffstat is any indication this shouldn't have a huge impact but it will have some. Most changes in the `str` and `path` module. A lot of the existing usages were in tests where ascii is expected. There are a number of other legit uses where the characters are known to be ascii.

Author: bors

src/libcollections/str.rs

@@ -1569,8 +1569,8 @@ mod tests {
         let n2: uint = v.len();
         assert_eq!(n1, n2);
         while i < n1 {
-            let a: u8 = s1.as_slice()[i];
-            let b: u8 = s2.as_slice()[i];
+            let a: u8 = s1.as_bytes()[i];
+            let b: u8 = s2.as_bytes()[i];
             debug!("{}", a);
             debug!("{}", b);
             assert_eq!(a, b);

src/libcollections/string.rs

@@ -222,7 +222,7 @@ impl String {
             return None
         }
 
-        let byte = self.as_slice()[len - 1];
+        let byte = self.as_bytes()[len - 1];
         self.vec.set_len(len - 1);
         Some(byte)
     }

src/libcore/str.rs

@@ -1743,7 +1743,7 @@ impl<'a> StrSlice<'a> for &'a str {
     fn lines_any(&self) -> AnyLines<'a> {
         self.lines().map(|line| {
             let l = line.len();
-            if l > 0 && line[l - 1] == '\r' as u8 { line.slice(0, l - 1) }
+            if l > 0 && line.as_bytes()[l - 1] == '\r' as u8 { line.slice(0, l - 1) }
             else { line }
         })
     }
@@ -1867,26 +1867,26 @@ impl<'a> StrSlice<'a> for &'a str {
     fn is_char_boundary(&self, index: uint) -> bool {
         if index == self.len() { return true; }
         if index > self.len() { return false; }
-        let b = self[index];
+        let b = self.as_bytes()[index];
         return b < 128u8 || b >= 192u8;
     }
 
     #[inline]
     fn char_range_at(&self, i: uint) -> CharRange {
-        if self[i] < 128u8 {
-            return CharRange {ch: self[i] as char, next: i + 1 };
+        if self.as_bytes()[i] < 128u8 {
+            return CharRange {ch: self.as_bytes()[i] as char, next: i + 1 };
         }
 
         // Multibyte case is a fn to allow char_range_at to inline cleanly
         fn multibyte_char_range_at(s: &str, i: uint) -> CharRange {
-            let mut val = s[i] as u32;
+            let mut val = s.as_bytes()[i] as u32;
             let w = UTF8_CHAR_WIDTH[val as uint] as uint;
             assert!((w != 0));
 
             val = utf8_first_byte!(val, w);
-            val = utf8_acc_cont_byte!(val, s[i + 1]);
-            if w > 2 { val = utf8_acc_cont_byte!(val, s[i + 2]); }
-            if w > 3 { val = utf8_acc_cont_byte!(val, s[i + 3]); }
+            val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 1]);
+            if w > 2 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 2]); }
+            if w > 3 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 3]); }
 
             return CharRange {ch: unsafe { mem::transmute(val) }, next: i + w};
         }
@@ -1899,23 +1899,25 @@ impl<'a> StrSlice<'a> for &'a str {
         let mut prev = start;
 
         prev = prev.saturating_sub(1);
-        if self[prev] < 128 { return CharRange{ch: self[prev] as char, next: prev} }
+        if self.as_bytes()[prev] < 128 {
+            return CharRange{ch: self.as_bytes()[prev] as char, next: prev}
+        }
 
         // Multibyte case is a fn to allow char_range_at_reverse to inline cleanly
         fn multibyte_char_range_at_reverse(s: &str, mut i: uint) -> CharRange {
             // while there is a previous byte == 10......
-            while i > 0 && s[i] & 192u8 == TAG_CONT_U8 {
+            while i > 0 && s.as_bytes()[i] & 192u8 == TAG_CONT_U8 {
                 i -= 1u;
             }
 
-            let mut val = s[i] as u32;
+            let mut val = s.as_bytes()[i] as u32;
             let w = UTF8_CHAR_WIDTH[val as uint] as uint;
             assert!((w != 0));
 
             val = utf8_first_byte!(val, w);
-            val = utf8_acc_cont_byte!(val, s[i + 1]);
-            if w > 2 { val = utf8_acc_cont_byte!(val, s[i + 2]); }
-            if w > 3 { val = utf8_acc_cont_byte!(val, s[i + 3]); }
+            val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 1]);
+            if w > 2 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 2]); }
+            if w > 3 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 3]); }
 
             return CharRange {ch: unsafe { mem::transmute(val) }, next: i};
         }

src/libgetopts/lib.rs

@@ -370,7 +370,7 @@ impl Matches {
 }
 
 fn is_arg(arg: &str) -> bool {
-    arg.len() > 1 && arg[0] == '-' as u8
+    arg.len() > 1 && arg.as_bytes()[0] == '-' as u8
 }
 
 fn find_opt(opts: &[Opt], nm: Name) -> Option<uint> {
@@ -553,7 +553,7 @@ pub fn getopts(args: &[String], optgrps: &[OptGroup]) -> Result {
         } else {
             let mut names;
             let mut i_arg = None;
-            if cur.as_slice()[1] == '-' as u8 {
+            if cur.as_bytes()[1] == '-' as u8 {
                 let tail = cur.as_slice().slice(2, curlen);
                 let tail_eq: Vec<&str> = tail.split('=').collect();
                 if tail_eq.len() <= 1 {

src/librustc/back/link.rs

@@ -657,8 +657,8 @@ pub fn sanitize(s: &str) -> String {
 
     // Underscore-qualify anything that didn't start as an ident.
     if result.len() > 0u &&
-        result.as_slice()[0] != '_' as u8 &&
-        ! char::is_XID_start(result.as_slice()[0] as char) {
+        result.as_bytes()[0] != '_' as u8 &&
+        ! char::is_XID_start(result.as_bytes()[0] as char) {
         return format!("_{}", result.as_slice());
     }
 
@@ -737,9 +737,9 @@ pub fn mangle_exported_name(ccx: &CrateContext, path: PathElems,
     let extra2 = id % EXTRA_CHARS.len();
     let id = id / EXTRA_CHARS.len();
     let extra3 = id % EXTRA_CHARS.len();
-    hash.push_char(EXTRA_CHARS[extra1] as char);
-    hash.push_char(EXTRA_CHARS[extra2] as char);
-    hash.push_char(EXTRA_CHARS[extra3] as char);
+    hash.push_char(EXTRA_CHARS.as_bytes()[extra1] as char);
+    hash.push_char(EXTRA_CHARS.as_bytes()[extra2] as char);
+    hash.push_char(EXTRA_CHARS.as_bytes()[extra3] as char);
 
     exported_name(path,
                   hash.as_slice(),

src/librustc/metadata/decoder.rs

@@ -181,7 +181,7 @@ fn item_sized(item: ebml::Doc) -> ast::Sized {
 fn item_method_sort(item: ebml::Doc) -> char {
     let mut ret = 'r';
     reader::tagged_docs(item, tag_item_trait_method_sort, |doc| {
-        ret = doc.as_str_slice()[0] as char;
+        ret = doc.as_str_slice().as_bytes()[0] as char;
         false
     });
     ret
@@ -757,13 +757,13 @@ fn get_explicit_self(item: ebml::Doc) -> ast::ExplicitSelf_ {
     let explicit_self_doc = reader::get_doc(item, tag_item_trait_method_explicit_self);
     let string = explicit_self_doc.as_str_slice();
 
-    let explicit_self_kind = string[0];
+    let explicit_self_kind = string.as_bytes()[0];
     match explicit_self_kind as char {
         's' => ast::SelfStatic,
         'v' => ast::SelfValue,
         '~' => ast::SelfUniq,
         // FIXME(#4846) expl. region
-        '&' => ast::SelfRegion(None, get_mutability(string[1])),
+        '&' => ast::SelfRegion(None, get_mutability(string.as_bytes()[1])),
         _ => fail!("unknown self type code: `{}`", explicit_self_kind as char)
     }
 }

src/librustc/middle/dead.rs

@@ -399,7 +399,7 @@ struct DeadVisitor<'a> {
 impl<'a> DeadVisitor<'a> {
     fn should_warn_about_field(&mut self, node: &ast::StructField_) -> bool {
         let (is_named, has_leading_underscore) = match node.ident() {
-            Some(ref ident) => (true, token::get_ident(*ident).get()[0] == ('_' as u8)),
+            Some(ref ident) => (true, token::get_ident(*ident).get().as_bytes()[0] == ('_' as u8)),
             _ => (false, false)
         };
         let field_type = ty::node_id_to_type(self.tcx, node.id);

src/librustc/middle/liveness.rs

@@ -1511,7 +1511,7 @@ impl<'a> Liveness<'a> {
 
     fn should_warn(&self, var: Variable) -> Option<String> {
         let name = self.ir.variable_name(var);
-        if name.len() == 0 || name.as_slice()[0] == ('_' as u8) {
+        if name.len() == 0 || name.as_bytes()[0] == ('_' as u8) {
             None
         } else {
             Some(name)

src/librustc/middle/mem_categorization.rs

@@ -126,7 +126,6 @@ pub enum FieldName {
 #[deriving(Clone, PartialEq, Eq, Hash)]
 pub enum ElementKind {
     VecElement,
-    StrElement,
     OtherElement,
 }
 
@@ -794,7 +793,7 @@ impl<'t,TYPER:Typer> MemCategorizationContext<'t,TYPER> {
         //! - `derefs`: the deref number to be used for
         //!   the implicit index deref, if any (see above)
 
-        let element_ty = match ty::index(base_cmt.ty) {
+        let element_ty = match ty::array_element_ty(base_cmt.ty) {
           Some(ref mt) => mt.ty,
           None => {
             self.tcx().sess.span_bug(
@@ -1137,9 +1136,6 @@ impl<'t,TYPER:Typer> MemCategorizationContext<'t,TYPER> {
           cat_interior(_, InteriorElement(VecElement)) => {
               "vec content".to_string()
           }
-          cat_interior(_, InteriorElement(StrElement)) => {
-              "str content".to_string()
-          }
           cat_interior(_, InteriorElement(OtherElement)) => {
               "indexed content".to_string()
           }
@@ -1320,7 +1316,6 @@ fn element_kind(t: ty::t) -> ElementKind {
         ty::ty_rptr(_, ty::mt{ty:ty, ..}) |
         ty::ty_uniq(ty) => match ty::get(ty).sty {
             ty::ty_vec(_, None) => VecElement,
-            ty::ty_str => StrElement,
             _ => OtherElement
         },
         ty::ty_vec(..) => VecElement,

src/librustc/middle/ty.rs

@@ -2563,6 +2563,21 @@ pub fn deref(t: t, explicit: bool) -> Option<mt> {
 
 // Returns the type of t[i]
 pub fn index(t: t) -> Option<mt> {
+    match get(t).sty {
+        ty_vec(mt, Some(_)) => Some(mt),
+        ty_ptr(mt{ty: t, ..}) | ty_rptr(_, mt{ty: t, ..}) |
+        ty_box(t) | ty_uniq(t) => match get(t).sty {
+            ty_vec(mt, None) => Some(mt),
+            _ => None,
+        },
+        _ => None
+    }
+}
+
+// Returns the type of elements contained within an 'array-like' type.
+// This is exactly the same as the above, except it supports strings,
+// which can't actually be indexed.
+pub fn array_element_ty(t: t) -> Option<mt> {
     match get(t).sty {
         ty_vec(mt, Some(_)) => Some(mt),
         ty_ptr(mt{ty: t, ..}) | ty_rptr(_, mt{ty: t, ..}) |