Merge 53691eec4e into 07d7fe2b33

core/asm/asm.go

@@ -0,0 +1,156 @@
+// Copyright 2017 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+// Package asm provides support for dealing with EVM assembly instructions (e.g., disassembling them).
+package asm
+
+import (
+	"encoding/hex"
+	"fmt"
+
+	"github.com/ethereum/go-ethereum/core/vm"
+)
+
+// Iterator for disassembled EVM instructions
+type instructionIterator struct {
+	code       []byte
+	pc         uint64
+	arg        []byte
+	op         vm.OpCode
+	error      error
+	started    bool
+	eofEnabled bool
+}
+
+// NewInstructionIterator creates a new instruction iterator.
+func NewInstructionIterator(code []byte) *instructionIterator {
+	it := new(instructionIterator)
+	it.code = code
+	return it
+}
+
+// NewEOFInstructionIterator creates a new instruction iterator for EOF-code.
+func NewEOFInstructionIterator(code []byte) *instructionIterator {
+	it := NewInstructionIterator(code)
+	it.eofEnabled = true
+	return it
+}
+
+// Next returns true if there is a next instruction and moves on.
+func (it *instructionIterator) Next() bool {
+	if it.error != nil || uint64(len(it.code)) <= it.pc {
+		// We previously reached an error or the end.
+		return false
+	}
+
+	if it.started {
+		// Since the iteration has been already started we move to the next instruction.
+		if it.arg != nil {
+			it.pc += uint64(len(it.arg))
+		}
+		it.pc++
+	} else {
+		// We start the iteration from the first instruction.
+		it.started = true
+	}
+
+	if uint64(len(it.code)) <= it.pc {
+		// We reached the end.
+		return false
+	}
+	it.op = vm.OpCode(it.code[it.pc])
+	var a int
+	if !it.eofEnabled { // Legacy code
+		if it.op.IsPush() {
+			a = int(it.op) - int(vm.PUSH0)
+		}
+	} else { // EOF code
+		if it.op == vm.RJUMPV {
+			// RJUMPV is unique as it has a variable sized operand. The total size is
+			// determined by the count byte which immediately follows RJUMPV.
+			maxIndex := int(it.code[it.pc+1])
+			a = (maxIndex+1)*2 + 1
+		} else {
+			a = vm.Immediates(it.op)
+		}
+	}
+	if a > 0 {
+		u := it.pc + 1 + uint64(a)
+		if uint64(len(it.code)) <= it.pc || uint64(len(it.code)) < u {
+			it.error = fmt.Errorf("incomplete instruction at %v", it.pc)
+			return false
+		}
+		it.arg = it.code[it.pc+1 : u]
+	} else {
+		it.arg = nil
+	}
+	return true
+}
+
+// Error returns any error that may have been encountered.
+func (it *instructionIterator) Error() error {
+	return it.error
+}
+
+// PC returns the PC of the current instruction.
+func (it *instructionIterator) PC() uint64 {
+	return it.pc
+}
+
+// Op returns the opcode of the current instruction.
+func (it *instructionIterator) Op() vm.OpCode {
+	return it.op
+}
+
+// Arg returns the argument of the current instruction.
+func (it *instructionIterator) Arg() []byte {
+	return it.arg
+}
+
+// PrintDisassembled pretty-print all disassembled EVM instructions to stdout.
+func PrintDisassembled(code string) error {
+	script, err := hex.DecodeString(code)
+	if err != nil {
+		return err
+	}
+	it := NewInstructionIterator(script)
+	for it.Next() {
+		if it.Arg() != nil && 0 < len(it.Arg()) {
+			fmt.Printf("%05x: %v %#x\n", it.PC(), it.Op(), it.Arg())
+		} else {
+			fmt.Printf("%05x: %v\n", it.PC(), it.Op())
+		}
+	}
+	return it.Error()
+}
+
+// Disassemble returns all disassembled EVM instructions in human-readable format.
+func Disassemble(script []byte) ([]string, error) {
+	instrs := make([]string, 0)
+
+	it := NewInstructionIterator(script)
+	for it.Next() {
+		if it.Arg() != nil && 0 < len(it.Arg()) {
+			instrs = append(instrs, fmt.Sprintf("%05x: %v %#x\n", it.PC(), it.Op(), it.Arg()))
+		} else {
+			instrs = append(instrs, fmt.Sprintf("%05x: %v\n", it.PC(), it.Op()))
+		}
+	}
+	if err := it.Error(); err != nil {
+		return nil, err
+	}
+	return instrs, nil
+}

core/asm/asm_test.go

@@ -0,0 +1,94 @@
+// Copyright 2017 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package asm
+
+import (
+	"encoding/hex"
+	"fmt"
+	"strings"
+	"testing"
+)
+
+// Tests disassembling instructions
+func TestInstructionIterator(t *testing.T) {
+	for i, tc := range []struct {
+		code       string
+		legacyWant string
+		eofWant    string
+	}{
+		{"", "", ""}, // empty
+		{"6100", `err: incomplete instruction at 0`, `err: incomplete instruction at 0`},
+		{"61000000", `
+00000: PUSH2 0x0000
+00003: STOP`, `
+00000: PUSH2 0x0000
+00003: STOP`},
+		{"5F00", `
+00000: PUSH0
+00001: STOP`, `
+00000: PUSH0
+00001: STOP`},
+		{"d1aabb00", `00000: DATALOADN
+00001: opcode 0xaa not defined
+00002: opcode 0xbb not defined
+00003: STOP`, `
+00000: DATALOADN 0xaabb
+00003: STOP`}, // DATALOADN(aabb),STOP
+		{"d1aa", `
+00000: DATALOADN
+00001: opcode 0xaa not defined`, "err: incomplete instruction at 0\n"}, // DATALOADN(aa) invalid
+		{"e20211223344556600", `
+00000: RJUMPV
+00001: MUL
+00002: GT
+00003: opcode 0x22 not defined
+00004: CALLER
+00005: DIFFICULTY
+00006: SSTORE
+err: incomplete instruction at 7`, `
+00000: RJUMPV 0x02112233445566
+00008: STOP`}, // RJUMPV( 6 bytes), STOP
+
+	} {
+		var (
+			code, _ = hex.DecodeString(tc.code)
+			legacy  = strings.TrimSpace(disassembly(NewInstructionIterator(code)))
+			eof     = strings.TrimSpace(disassembly(NewEOFInstructionIterator(code)))
+		)
+		if want := strings.TrimSpace(tc.legacyWant); legacy != want {
+			t.Errorf("test %d: wrong (legacy) output. have:\n%q\nwant:\n%q\n", i, legacy, want)
+		}
+		if want := strings.TrimSpace(tc.eofWant); eof != want {
+			t.Errorf("test %d: wrong (eof) output. have:\n%q\nwant:\n%q\n", i, eof, want)
+		}
+	}
+}
+
+func disassembly(it *instructionIterator) string {
+	var out = new(strings.Builder)
+	for it.Next() {
+		if it.Arg() != nil && 0 < len(it.Arg()) {
+			fmt.Fprintf(out, "%05x: %v %#x\n", it.PC(), it.Op(), it.Arg())
+		} else {
+			fmt.Fprintf(out, "%05x: %v\n", it.PC(), it.Op())
+		}
+	}
+	if err := it.Error(); err != nil {
+		fmt.Fprintf(out, "err: %v\n", err)
+	}
+	return out.String()
+}

core/asm/compiler.go

@@ -0,0 +1,292 @@
+// Copyright 2017 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package asm
+
+import (
+	"encoding/hex"
+	"errors"
+	"fmt"
+	"math/big"
+	"os"
+	"strings"
+
+	"github.com/ethereum/go-ethereum/common/math"
+	"github.com/ethereum/go-ethereum/core/vm"
+)
+
+// Compiler contains information about the parsed source
+// and holds the tokens for the program.
+type Compiler struct {
+	tokens []token
+	out    []byte
+
+	labels map[string]int
+
+	pc, pos int
+
+	debug bool
+}
+
+// NewCompiler returns a new allocated compiler.
+func NewCompiler(debug bool) *Compiler {
+	return &Compiler{
+		labels: make(map[string]int),
+		debug:  debug,
+	}
+}
+
+// Feed feeds tokens into ch and are interpreted by
+// the compiler.
+//
+// feed is the first pass in the compile stage as it collects the used labels in the
+// program and keeps a program counter which is used to determine the locations of the
+// jump dests. The labels can than be used in the second stage to push labels and
+// determine the right position.
+func (c *Compiler) Feed(ch <-chan token) {
+	var prev token
+	for i := range ch {
+		switch i.typ {
+		case number:
+			num := math.MustParseBig256(i.text).Bytes()
+			if len(num) == 0 {
+				num = []byte{0}
+			}
+			c.pc += len(num)
+		case stringValue:
+			c.pc += len(i.text) - 2
+		case element:
+			c.pc++
+		case labelDef:
+			c.labels[i.text] = c.pc
+			c.pc++
+		case label:
+			c.pc += 4
+			if prev.typ == element && isJump(prev.text) {
+				c.pc++
+			}
+		}
+		c.tokens = append(c.tokens, i)
+		prev = i
+	}
+	if c.debug {
+		fmt.Fprintln(os.Stderr, "found", len(c.labels), "labels")
+	}
+}
+
+// Compile compiles the current tokens and returns a binary string that can be interpreted
+// by the EVM and an error if it failed.
+//
+// compile is the second stage in the compile phase which compiles the tokens to EVM
+// instructions.
+func (c *Compiler) Compile() (string, []error) {
+	var errors []error
+	// continue looping over the tokens until
+	// the stack has been exhausted.
+	for c.pos < len(c.tokens) {
+		if err := c.compileLine(); err != nil {
+			errors = append(errors, err)
+		}
+	}
+
+	// turn the binary to hex
+	h := hex.EncodeToString(c.out)
+	return h, errors
+}
+
+// next returns the next token and increments the
+// position.
+func (c *Compiler) next() token {
+	token := c.tokens[c.pos]
+	c.pos++
+	return token
+}
+
+// compileLine compiles a single line instruction e.g.
+// "push 1", "jump @label".
+func (c *Compiler) compileLine() error {
+	n := c.next()
+	if n.typ != lineStart {
+		return compileErr(n, n.typ.String(), lineStart.String())
+	}
+
+	lvalue := c.next()
+	switch lvalue.typ {
+	case eof:
+		return nil
+	case element:
+		if err := c.compileElement(lvalue); err != nil {
+			return err
+		}
+	case labelDef:
+		c.compileLabel()
+	case lineEnd:
+		return nil
+	default:
+		return compileErr(lvalue, lvalue.text, fmt.Sprintf("%v or %v", labelDef, element))
+	}
+
+	if n := c.next(); n.typ != lineEnd {
+		return compileErr(n, n.text, lineEnd.String())
+	}
+
+	return nil
+}
+
+// parseNumber compiles the number to bytes
+func parseNumber(tok token) ([]byte, error) {
+	if tok.typ != number {
+		panic("parseNumber of non-number token")
+	}
+	num, ok := math.ParseBig256(tok.text)
+	if !ok {
+		return nil, errors.New("invalid number")
+	}
+	bytes := num.Bytes()
+	if len(bytes) == 0 {
+		bytes = []byte{0}
+	}
+	return bytes, nil
+}
+
+// compileElement compiles the element (push & label or both)
+// to a binary representation and may error if incorrect statements
+// where fed.
+func (c *Compiler) compileElement(element token) error {
+	switch {
+	case isJump(element.text):
+		return c.compileJump(element.text)
+	case isPush(element.text):
+		return c.compilePush()
+	default:
+		c.outputOpcode(toBinary(element.text))
+		return nil
+	}
+}
+
+func (c *Compiler) compileJump(jumpType string) error {
+	rvalue := c.next()
+	switch rvalue.typ {
+	case number:
+		numBytes, err := parseNumber(rvalue)
+		if err != nil {
+			return err
+		}
+		c.outputBytes(numBytes)
+
+	case stringValue:
+		// strings are quoted, remove them.
+		str := rvalue.text[1 : len(rvalue.text)-2]
+		c.outputBytes([]byte(str))
+
+	case label:
+		c.outputOpcode(vm.PUSH4)
+		pos := big.NewInt(int64(c.labels[rvalue.text])).Bytes()
+		pos = append(make([]byte, 4-len(pos)), pos...)
+		c.outputBytes(pos)
+
+	case lineEnd:
+		// push without argument is supported, it just takes the destination from the stack.
+		c.pos--
+
+	default:
+		return compileErr(rvalue, rvalue.text, "number, string or label")
+	}
+	// push the operation
+	c.outputOpcode(toBinary(jumpType))
+	return nil
+}
+
+func (c *Compiler) compilePush() error {
+	// handle pushes. pushes are read from left to right.
+	var value []byte
+	rvalue := c.next()
+	switch rvalue.typ {
+	case number:
+		value = math.MustParseBig256(rvalue.text).Bytes()
+		if len(value) == 0 {
+			value = []byte{0}
+		}
+	case stringValue:
+		value = []byte(rvalue.text[1 : len(rvalue.text)-1])
+	case label:
+		value = big.NewInt(int64(c.labels[rvalue.text])).Bytes()
+		value = append(make([]byte, 4-len(value)), value...)
+	default:
+		return compileErr(rvalue, rvalue.text, "number, string or label")
+	}
+	if len(value) > 32 {
+		return fmt.Errorf("%d: string or number size > 32 bytes", rvalue.lineno+1)
+	}
+	c.outputOpcode(vm.OpCode(int(vm.PUSH1) - 1 + len(value)))
+	c.outputBytes(value)
+	return nil
+}
+
+// compileLabel pushes a jumpdest to the binary slice.
+func (c *Compiler) compileLabel() {
+	c.outputOpcode(vm.JUMPDEST)
+}
+
+func (c *Compiler) outputOpcode(op vm.OpCode) {
+	if c.debug {
+		fmt.Printf("%d: %v\n", len(c.out), op)
+	}
+	c.out = append(c.out, byte(op))
+}
+
+// output pushes the value v to the binary stack.
+func (c *Compiler) outputBytes(b []byte) {
+	if c.debug {
+		fmt.Printf("%d: %x\n", len(c.out), b)
+	}
+	c.out = append(c.out, b...)
+}
+
+// isPush returns whether the string op is either any of
+// push(N).
+func isPush(op string) bool {
+	return strings.EqualFold(op, "PUSH")
+}
+
+// isJump returns whether the string op is jump(i)
+func isJump(op string) bool {
+	return strings.EqualFold(op, "JUMPI") || strings.EqualFold(op, "JUMP")
+}
+
+// toBinary converts text to a vm.OpCode
+func toBinary(text string) vm.OpCode {
+	return vm.StringToOp(strings.ToUpper(text))
+}
+
+type compileError struct {
+	got  string
+	want string
+
+	lineno int
+}
+
+func (err compileError) Error() string {
+	return fmt.Sprintf("%d: syntax error: unexpected %v, expected %v", err.lineno, err.got, err.want)
+}
+
+func compileErr(c token, got, want string) error {
+	return compileError{
+		got:    got,
+		want:   want,
+		lineno: c.lineno + 1,
+	}
+}

core/asm/compiler_test.go

@@ -0,0 +1,79 @@
+// Copyright 2019 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package asm
+
+import (
+	"testing"
+)
+
+func TestCompiler(t *testing.T) {
+	tests := []struct {
+		input, output string
+	}{
+		{
+			input: `
+	GAS
+	label:
+	PUSH @label
+`,
+			output: "5a5b6300000001",
+		},
+		{
+			input: `
+	PUSH @label
+	label:
+`,
+			output: "63000000055b",
+		},
+		{
+			input: `
+	PUSH @label
+	JUMP
+	label:
+`,
+			output: "6300000006565b",
+		},
+		{
+			input: `
+	JUMP @label
+	label:
+`,
+			output: "6300000006565b",
+		},
+		{
+			input: `
+	JUMP @label
+label: ;; comment
+	ADD ;; comment
+`,
+			output: "6300000006565b01",
+		},
+	}
+	for _, test := range tests {
+		ch := Lex([]byte(test.input), false)
+		c := NewCompiler(false)
+		c.Feed(ch)
+		output, err := c.Compile()
+		if len(err) != 0 {
+			t.Errorf("compile error: %v\ninput: %s", err, test.input)
+			continue
+		}
+		if output != test.output {
+			t.Errorf("incorrect output\ninput: %sgot:  %s\nwant: %s\n", test.input, output, test.output)
+		}
+	}
+}

core/asm/lex_test.go

@@ -0,0 +1,93 @@
+// Copyright 2017 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package asm
+
+import (
+	"reflect"
+	"testing"
+)
+
+func lexAll(src string) []token {
+	ch := Lex([]byte(src), false)
+
+	var tokens []token
+	for i := range ch {
+		tokens = append(tokens, i)
+	}
+	return tokens
+}
+
+func TestLexer(t *testing.T) {
+	tests := []struct {
+		input  string
+		tokens []token
+	}{
+		{
+			input:  ";; this is a comment",
+			tokens: []token{{typ: lineStart}, {typ: eof}},
+		},
+		{
+			input:  "0x12345678",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "0x12345678"}, {typ: eof}},
+		},
+		{
+			input:  "0x123ggg",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "0x123"}, {typ: element, text: "ggg"}, {typ: eof}},
+		},
+		{
+			input:  "12345678",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "12345678"}, {typ: eof}},
+		},
+		{
+			input:  "123abc",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "123"}, {typ: element, text: "abc"}, {typ: eof}},
+		},
+		{
+			input:  "0123abc",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "0123"}, {typ: element, text: "abc"}, {typ: eof}},
+		},
+		{
+			input:  "00123abc",
+			tokens: []token{{typ: lineStart}, {typ: number, text: "00123"}, {typ: element, text: "abc"}, {typ: eof}},
+		},
+		{
+			input:  "@foo",
+			tokens: []token{{typ: lineStart}, {typ: label, text: "foo"}, {typ: eof}},
+		},
+		{
+			input:  "@label123",
+			tokens: []token{{typ: lineStart}, {typ: label, text: "label123"}, {typ: eof}},
+		},
+		// Comment after label
+		{
+			input:  "@label123 ;; comment",
+			tokens: []token{{typ: lineStart}, {typ: label, text: "label123"}, {typ: eof}},
+		},
+		// Comment after instruction
+		{
+			input:  "push 3 ;; comment\nadd",
+			tokens: []token{{typ: lineStart}, {typ: element, text: "push"}, {typ: number, text: "3"}, {typ: lineEnd, text: "\n"}, {typ: lineStart, lineno: 1}, {typ: element, lineno: 1, text: "add"}, {typ: eof, lineno: 1}},
+		},
+	}
+
+	for _, test := range tests {
+		tokens := lexAll(test.input)
+		if !reflect.DeepEqual(tokens, test.tokens) {
+			t.Errorf("input %q\ngot:  %+v\nwant: %+v", test.input, tokens, test.tokens)
+		}
+	}
+}

core/asm/lexer.go

@@ -0,0 +1,275 @@
+// Copyright 2017 The go-ethereum Authors
+// This file is part of the go-ethereum library.
+//
+// The go-ethereum library is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Lesser General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// The go-ethereum library is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+// GNU Lesser General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public License
+// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
+
+package asm
+
+import (
+	"fmt"
+	"os"
+	"strings"
+	"unicode"
+	"unicode/utf8"
+)
+
+// stateFn is used through the lifetime of the
+// lexer to parse the different values at the
+// current state.
+type stateFn func(*lexer) stateFn
+
+// token is emitted when the lexer has discovered
+// a new parsable token. These are delivered over
+// the tokens channels of the lexer
+type token struct {
+	typ    tokenType
+	lineno int
+	text   string
+}
+
+// tokenType are the different types the lexer
+// is able to parse and return.
+type tokenType int
+
+//go:generate go run golang.org/x/tools/cmd/stringer -type tokenType
+
+const (
+	eof              tokenType = iota // end of file
+	lineStart                         // emitted when a line starts
+	lineEnd                           // emitted when a line ends
+	invalidStatement                  // any invalid statement
+	element                           // any element during element parsing
+	label                             // label is emitted when a label is found
+	labelDef                          // label definition is emitted when a new label is found
+	number                            // number is emitted when a number is found
+	stringValue                       // stringValue is emitted when a string has been found
+)
+
+const (
+	decimalNumbers = "1234567890"                                           // characters representing any decimal number
+	hexNumbers     = decimalNumbers + "aAbBcCdDeEfF"                        // characters representing any hexadecimal
+	alpha          = "abcdefghijklmnopqrstuwvxyzABCDEFGHIJKLMNOPQRSTUWVXYZ" // characters representing alphanumeric
+)
+
+// lexer is the basic construct for parsing
+// source code and turning them in to tokens.
+// Tokens are interpreted by the compiler.
+type lexer struct {
+	input string // input contains the source code of the program
+
+	tokens chan token // tokens is used to deliver tokens to the listener
+	state  stateFn    // the current state function
+
+	lineno            int // current line number in the source file
+	start, pos, width int // positions for lexing and returning value
+
+	debug bool // flag for triggering debug output
+}
+
+// Lex lexes the program by name with the given source. It returns a
+// channel on which the tokens are delivered.
+func Lex(source []byte, debug bool) <-chan token {
+	ch := make(chan token)
+	l := &lexer{
+		input:  string(source),
+		tokens: ch,
+		state:  lexLine,
+		debug:  debug,
+	}
+	go func() {
+		l.emit(lineStart)
+		for l.state != nil {
+			l.state = l.state(l)
+		}
+		l.emit(eof)
+		close(l.tokens)
+	}()
+
+	return ch
+}
+
+// next returns the next rune in the program's source.
+func (l *lexer) next() (rune rune) {
+	if l.pos >= len(l.input) {
+		l.width = 0
+		return 0
+	}
+	rune, l.width = utf8.DecodeRuneInString(l.input[l.pos:])
+	l.pos += l.width
+	return rune
+}
+
+// backup backsup the last parsed element (multi-character)
+func (l *lexer) backup() {
+	l.pos -= l.width
+}
+
+// peek returns the next rune but does not advance the seeker
+func (l *lexer) peek() rune {
+	r := l.next()
+	l.backup()
+	return r
+}
+
+// ignore advances the seeker and ignores the value
+func (l *lexer) ignore() {
+	l.start = l.pos
+}
+
+// accept checks whether the given input matches the next rune
+func (l *lexer) accept(valid string) bool {
+	if strings.ContainsRune(valid, l.next()) {
+		return true
+	}
+
+	l.backup()
+
+	return false
+}
+
+// acceptRun will continue to advance the seeker until valid
+// can no longer be met.
+func (l *lexer) acceptRun(valid string) {
+	for strings.ContainsRune(valid, l.next()) {
+	}
+	l.backup()
+}
+
+// acceptRunUntil is the inverse of acceptRun and will continue
+// to advance the seeker until the rune has been found.
+func (l *lexer) acceptRunUntil(until rune) bool {
+	// Continues running until a rune is found
+	for i := l.next(); !strings.ContainsRune(string(until), i); i = l.next() {
+		if i == 0 {
+			return false
+		}
+	}
+
+	return true
+}
+
+// blob returns the current value
+func (l *lexer) blob() string {
+	return l.input[l.start:l.pos]
+}
+
+// Emits a new token on to token channel for processing
+func (l *lexer) emit(t tokenType) {
+	token := token{t, l.lineno, l.blob()}
+
+	if l.debug {
+		fmt.Fprintf(os.Stderr, "%04d: (%-20v) %s\n", token.lineno, token.typ, token.text)
+	}
+
+	l.tokens <- token
+	l.start = l.pos
+}
+
+// lexLine is state function for lexing lines
+func lexLine(l *lexer) stateFn {
+	for {
+		switch r := l.next(); {
+		case r == '\n':
+			l.emit(lineEnd)
+			l.ignore()
+			l.lineno++
+			l.emit(lineStart)
+		case r == ';' && l.peek() == ';':
+			return lexComment
+		case isSpace(r):
+			l.ignore()
+		case isLetter(r) || r == '_':
+			return lexElement
+		case isNumber(r):
+			return lexNumber
+		case r == '@':
+			l.ignore()
+			return lexLabel
+		case r == '"':
+			return lexInsideString
+		default:
+			return nil
+		}
+	}
+}
+
+// lexComment parses the current position until the end
+// of the line and discards the text.
+func lexComment(l *lexer) stateFn {
+	l.acceptRunUntil('\n')
+	l.backup()
+	l.ignore()
+
+	return lexLine
+}
+
+// lexLabel parses the current label, emits and returns
+// the lex text state function to advance the parsing
+// process.
+func lexLabel(l *lexer) stateFn {
+	l.acceptRun(alpha + "_" + decimalNumbers)
+
+	l.emit(label)
+
+	return lexLine
+}
+
+// lexInsideString lexes the inside of a string until
+// the state function finds the closing quote.
+// It returns the lex text state function.
+func lexInsideString(l *lexer) stateFn {
+	if l.acceptRunUntil('"') {
+		l.emit(stringValue)
+	}
+
+	return lexLine
+}
+
+func lexNumber(l *lexer) stateFn {
+	acceptance := decimalNumbers
+	if l.accept("xX") {
+		acceptance = hexNumbers
+	}
+	l.acceptRun(acceptance)
+
+	l.emit(number)
+
+	return lexLine
+}
+
+func lexElement(l *lexer) stateFn {
+	l.acceptRun(alpha + "_" + decimalNumbers)
+
+	if l.peek() == ':' {
+		l.emit(labelDef)
+
+		l.accept(":")
+		l.ignore()
+	} else {
+		l.emit(element)
+	}
+	return lexLine
+}
+
+func isLetter(t rune) bool {
+	return unicode.IsLetter(t)
+}
+
+func isSpace(t rune) bool {
+	return unicode.IsSpace(t)
+}
+
+func isNumber(t rune) bool {
+	return unicode.IsNumber(t)
+}

core/asm/tokentype_string.go

@@ -0,0 +1,31 @@
+// Code generated by "stringer -type tokenType"; DO NOT EDIT.
+
+package asm
+
+import "strconv"
+
+func _() {
+	// An "invalid array index" compiler error signifies that the constant values have changed.
+	// Re-run the stringer command to generate them again.
+	var x [1]struct{}
+	_ = x[eof-0]
+	_ = x[lineStart-1]
+	_ = x[lineEnd-2]
+	_ = x[invalidStatement-3]
+	_ = x[element-4]
+	_ = x[label-5]
+	_ = x[labelDef-6]
+	_ = x[number-7]
+	_ = x[stringValue-8]
+}
+
+const _tokenType_name = "eoflineStartlineEndinvalidStatementelementlabellabelDefnumberstringValue"
+
+var _tokenType_index = [...]uint8{0, 3, 12, 19, 35, 42, 47, 55, 61, 72}
+
+func (i tokenType) String() string {
+	if i < 0 || i >= tokenType(len(_tokenType_index)-1) {
+		return "tokenType(" + strconv.FormatInt(int64(i), 10) + ")"
+	}
+	return _tokenType_name[_tokenType_index[i]:_tokenType_index[i+1]]
+}

core/vm/runtime/runtime_test.go

@@ -29,6 +29,7 @@ import (
 	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/consensus"
 	"github.com/ethereum/go-ethereum/core"
+	"github.com/ethereum/go-ethereum/core/asm"
 	"github.com/ethereum/go-ethereum/core/state"
 	"github.com/ethereum/go-ethereum/core/tracing"
 	"github.com/ethereum/go-ethereum/core/types"
@@ -507,9 +508,21 @@ func TestEip2929Cases(t *testing.T) {
 	t.Skip("Test only useful for generating documentation")
 	id := 1
 	prettyPrint := func(comment string, code []byte) {
+		instrs := make([]string, 0)
+		it := asm.NewInstructionIterator(code)
+		for it.Next() {
+			if it.Arg() != nil && 0 < len(it.Arg()) {
+				instrs = append(instrs, fmt.Sprintf("%v %#x", it.Op(), it.Arg()))
+			} else {
+				instrs = append(instrs, fmt.Sprintf("%v", it.Op()))
+			}
+		}
+		ops := strings.Join(instrs, ", ")
 		fmt.Printf("### Case %d\n\n", id)
 		id++
-		fmt.Printf("%v\n\nBytecode: \n```\n%#x\n```\n", comment, code)
+		fmt.Printf("%v\n\nBytecode: \n```\n%#x\n```\nOperations: \n```\n%v\n```\n\n",
+			comment,
+			code, ops)
 		Execute(code, nil, &Config{
 			EVMConfig: vm.Config{
 				Tracer:    logger.NewMarkdownLogger(nil, os.Stdout).Hooks(),