directives that cmd/link is not sophisticated enough to derive from _all.o
on its own. Similarly, the _all.o uses dynamic references to real
system object code because cmd/link is not sophisticated enough to process
the real code.

The main benefits of this system are that cmd/link remains relatively simple
(it does not need to implement a complete ELF and Mach-O linker) and

The built-in function <code>complex</code> constructs a complex
value from a floating-point real and imaginary part, while
<code>real</code> and <code>imag</code>
extract the real and imaginary parts of a complex value.
</p>

<pre class="grammar">
complex(realPart, imaginaryPart floatT) complexT
real(complexT) floatT
imag(complexT) floatT
</pre>

<p>

pkg go/constant, func MakeString(string) Value
pkg go/constant, func MakeUint64(uint64) Value
pkg go/constant, func MakeUnknown() Value
pkg go/constant, func Num(Value) Value
pkg go/constant, func Real(Value) Value
pkg go/constant, func Shift(Value, token.Token, uint) Value
pkg go/constant, func Sign(Value) int
pkg go/constant, func StringVal(Value) string
pkg go/constant, func Uint64Val(Value) (uint64, bool)

	//
	// When the function argument is a conversion to unsafe.Pointer,
	// we unwrap the conversion before checking the pointer,
	// and then wrap again when calling C.f. This lets us check
	// the real type of the pointer in some cases. See issue #25941.
	//
	// When the call to C.f is deferred, we use an additional function
	// literal to evaluate the arguments at the right time.
	//    defer func() func() {

// floating-point values. The real and imaginary parts must be of the same
// size, either float32 or float64 (or assignable to them), and the return
// value will be the corresponding complex type (complex64 for float32,
// complex128 for float64).
func complex(r, i FloatType) ComplexType

// The real built-in function returns the real part of the complex number c.

The built-in function <code>complex</code> constructs a complex
value from a floating-point real and imaginary part, while
<code>real</code> and <code>imag</code>
extract the real and imaginary parts of a complex value.
</p>

<pre class="grammar">
complex(realPart, imaginaryPart floatT) complexT
real(complexT) floatT
imag(complexT) floatT
</pre>

<p>

	JMP	(X5)					// 67800200
	JMP	4(X5)					// 67804200

	// CALL and JMP to symbol are encoded as JAL (using LR or ZERO
	// respectively), with a R_RISCV_JAL relocation. The linker resolves
	// the real address and updates the immediate, using a trampoline in
	// the case where the address is not directly reachable.
	CALL	asmtest(SB)				// ef000000
	JMP	asmtest(SB)				// 6f000000

	// Branch pseudo-instructions

	// FullRune distinguishes correctly between erroneous and incomplete encodings.
	if !atEOF && !utf8.FullRune(data) {
		// Incomplete; get more bytes.
		return 0, nil, nil
	}

	// We have a real UTF-8 encoding error. Return a properly encoded error rune
	// but advance only one byte. This matches the behavior of a range loop over
	// an incorrectly encoded string.
	return 1, errorRune, nil
}

are predefined and refer to registers.
The exact set depends on the architecture.
</p>

<p>
There are four predeclared symbols that refer to pseudo-registers.
These are not real registers, but rather virtual registers maintained by
the toolchain, such as a frame pointer.
The set of pseudo-registers is the same for all architectures:
</p>

<ul>

<li>

// Copyright 2018 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.

// It's going to be hard to include a whole real JVM to test this.
// So we'll simulate a really easy JVM using just the parts we need.

// This is the relevant part of jni.h.

// On Android NDK16, jobject is defined like this in C and C++
typedef void* jobject;

typedef jobject jclass;