Consider all zero-sized data structures to be voidish, bypassing some "quirky" parts of LLVM (see e.g. LLVM bug 9900) but also generating better code

Major changes:

- Define temporary scopes in a syntax-based way that basically defaults
  to the innermost statement or conditional block, except for in
  a `let` initializer, where we default to the innermost block. Rules
  are documented in the code, but not in the manual (yet).
  See new test run-pass/cleanup-value-scopes.rs for examples.
- Refactors Datum to better define cleanup roles.
- Refactor cleanup scopes to not be tied to basic blocks, permitting
  us to have a very large number of scopes (one per AST node).
- Introduce nascent documentation in trans/doc.rs covering datums and
  cleanup in a more comprehensive way.

Fixes #7411, #10615.

stack-allocate function contexts.

There's no need for the restrictions of a closure with the above methods.

and librustpkg.

This adds an other ABI option which allows a custom selection over the target
architecture and OS. The only current candidate for this change is that kernel32
on win32 uses stdcall, but on win64 it uses the cdecl calling convention.
Otherwise everywhere else this is defined as using the Cdecl calling convention.

cc #10049
Closes #8774

Who doesn't like a massive renaming?

This commit resumes management of the stack boundaries and limits when switching
between tasks. This additionally leverages the __morestack function to run code
on "stack overflow". The current behavior is to abort the process, but this is
probably not the best behavior in the long term (for deails, see the comment I
wrote up in the stack exhaustion routine).

I borrow some ideas from clang's ABIInfo.h and TargetInfo.cpp.
LLVMType is replaced with ArgType, which is similar to clang's ABIArgInfo,
and I also merge attrs of FnType into it.

Now ABI implementation doesn't need to insert hidden return pointer
to arg_tys of FnType. Instead it is handled in foreign.rs.

This change also fixes LLVM assertion failure when compiling MIPS target.

This fixes two existing bugs along the way:

* The `transmute` intrinsic did not correctly handle casts of immediate
  aggregates like newtype structs and tuples.
* The code for calling foreign functions used the wrong type to create
  an `alloca` temporary

    enum Foo { A, B }
    fn foo() -> Foo { A }

Before:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3foo18hbedc642d5d9cf5aag4v0.0E(%enum.Foo* noalias nocapture sret, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #0 {
    "function top level":
      %2 = getelementptr inbounds %enum.Foo* %0, i64 0, i32 0
      store i64 0, i64* %2, align 8
      ret void
    }

After:

    ; Function Attrs: nounwind readnone uwtable
    define %enum.Foo @_ZN3foo18hbedc642d5d9cf5aag4v0.0E({ i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #0 {
    "function top level":
      ret %enum.Foo zeroinitializer
    }

    fn foo() -> (u32, u8, u8, u8, u8) {
        (4, 5, 6, 7, 8)
    }

Before:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3foo18hbb616262f874f8daf4v0.0E({ i32, i8, i8, i8, i8 }* noalias nocapture sret, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #0 {
    "function top level":
      %2 = getelementptr inbounds { i32, i8, i8, i8, i8 }* %0, i64 0, i32 0
      store i32 4, i32* %2, align 4
      %3 = getelementptr inbounds { i32, i8, i8, i8, i8 }* %0, i64 0, i32 1
      store i8 5, i8* %3, align 4
      %4 = getelementptr inbounds { i32, i8, i8, i8, i8 }* %0, i64 0, i32 2
      store i8 6, i8* %4, align 1
      %5 = getelementptr inbounds { i32, i8, i8, i8, i8 }* %0, i64 0, i32 3
      store i8 7, i8* %5, align 2
      %6 = getelementptr inbounds { i32, i8, i8, i8, i8 }* %0, i64 0, i32 4
      store i8 8, i8* %6, align 1
      ret void
    }

After:

    ; Function Attrs: nounwind readnone uwtable
    define { i32, i8, i8, i8, i8 } @_ZN3foo18hbb616262f874f8daf4v0.0E({ i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #0 {
    "function top level":
      ret { i32, i8, i8, i8, i8 } { i32 4, i8 5, i8 6, i8 7, i8 8 }
    }

This brings Rust in line with how `clang` handles return pointers.

Example:

    pub fn bar() -> [uint, .. 8] {
        let a = [0, .. 8];
        a
    }

Before:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3bar17ha4635c6f704bfa334v0.0E([8 x i64]* nocapture, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #1 {
    "function top level":
      %a = alloca [8 x i64], align 8
      %2 = bitcast [8 x i64]* %a to i8*
      call void @llvm.memset.p0i8.i64(i8* %2, i8 0, i64 64, i32 8, i1 false)
      %3 = bitcast [8 x i64]* %0 to i8*
      call void @llvm.memcpy.p0i8.p0i8.i64(i8* %3, i8* %2, i64 64, i32 8, i1 false)
      ret void
    }

After:

    ; Function Attrs: nounwind uwtable
    define void @_ZN3bar17ha4635c6f704bfa334v0.0E([8 x i64]* noalias nocapture sret, { i64, %tydesc*, i8*, i8*, i8 }* nocapture readnone) #1 {
    "function top level":
      %2 = bitcast [8 x i64]* %0 to i8*
      call void @llvm.memset.p0i8.i64(i8* %2, i8 0, i64 64, i32 8, i1 false)
      ret void
    }

Closes #9072
Closes #7298
Closes #9154

The `noalias` attributes were being set only on function definitions,
not on all declarations. This is harmless for `noalias`, but prevented
some optimization opportunities and is *not* harmless for other
attributes like `sret` with ABI implications.

Closes #9104

Beforehand it was assumed that the standard cdecl abi was used for all extern
fns of extern crates, but this reads the abi of the extern fn type and declares
the function in the local crate with the appropriate type.

Since function pointers do not carry along the function attributes with
them in the type, this needs to be set on the call instruction itself.

Closes #9152

cc #3678