From 5121164cf9098137ff2ee4227c35cefa9572d39b Mon Sep 17 00:00:00 2001
From: Michael McLoughlin <mmcloughlin@gmail.com>
Date: Tue, 7 Jan 2020 15:39:12 -0500
Subject: [PATCH] loader for asmdb

Updates #99
---
 internal/asmdb/doc.go        |   2 +
 internal/asmdb/parse.go      | 117 +++++++++++++++++++++++++++++++++++
 internal/asmdb/parse_test.go |  13 ++++
 internal/asmdb/testdata      |   1 +
 4 files changed, 133 insertions(+)
 create mode 100644 internal/asmdb/doc.go
 create mode 100644 internal/asmdb/parse.go
 create mode 100644 internal/asmdb/parse_test.go
 create mode 120000 internal/asmdb/testdata

diff --git a/internal/asmdb/doc.go b/internal/asmdb/doc.go
new file mode 100644
index 00000000..afdbd62a
--- /dev/null
+++ b/internal/asmdb/doc.go
@@ -0,0 +1,2 @@
+// Package asmdb provides access to the asmjit opcodes database.
+package asmdb
diff --git a/internal/asmdb/parse.go b/internal/asmdb/parse.go
new file mode 100644
index 00000000..6244c195
--- /dev/null
+++ b/internal/asmdb/parse.go
@@ -0,0 +1,117 @@
+package asmdb
+
+import (
+	"bufio"
+	"encoding/json"
+	"io"
+	"os"
+	"strings"
+)
+
+// Extension specifies an architecture extension.
+type Extension struct {
+	Name string `json:"name"`
+}
+
+// Attribute is specifies an instruction attribute.
+type Attribute struct {
+	Name string `json:"name"`
+	Type string `json:"type"`
+	Doc  string `json:"doc"`
+}
+
+// SpecialReg specifies a special register or flag.
+type SpecialReg struct {
+	Name  string `json:"name"`
+	Group string `json:"group"`
+	Doc   string `json:"doc"`
+}
+
+// Shortcut specifies a shorthand for a collection of attributes.
+type Shortcut struct {
+	Name   string `json:"name"`
+	Expand string `json:"expand"`
+}
+
+// Register defines a class of registers.
+type Register struct {
+	Kind  string   `json:"kind"`
+	Any   string   `json:"any"`
+	Names []string `json:"names"`
+}
+
+// Raw is an asmdb instruction database in its unprocessed form.
+type Raw struct {
+	Architectures []string            `json:"architectures"`
+	Extensions    []Extension         `json:"extensions"`
+	Attributes    []Attribute         `json:"attributes"`
+	SpecialRegs   []SpecialReg        `json:"specialRegs"`
+	Shortcuts     []Shortcut          `json:"shortcuts"`
+	Registers     map[string]Register `json:"registers"`
+	Instructions  [][]string          `json:"instructions"`
+}
+
+// ParseRaw parses the asmdb x86data.js file as a raw unprocessed data structure.
+func ParseRaw(r io.Reader) (*Raw, error) {
+	// Extract the JSON blob from the javascript file.
+	b, err := extractjson(r)
+	if err != nil {
+		return nil, err
+	}
+
+	// Parse JSON.
+	db := &Raw{}
+	if err := json.Unmarshal(b, db); err != nil {
+		return nil, err
+	}
+
+	return db, nil
+}
+
+// ParseRawFile parses an asmdb x86data.js file.
+func ParseRawFile(filename string) (*Raw, error) {
+	f, err := os.Open(filename)
+	if err != nil {
+		return nil, err
+	}
+	defer f.Close()
+	return ParseRaw(f)
+}
+
+// extractjson extracts the pure JSON part of the x86data.js file.
+func extractjson(r io.Reader) ([]byte, error) {
+	dividers := []string{
+		`// ${JSON:BEGIN}`,
+		`// ${JSON:END}`,
+		`/*`,
+		`*/`,
+	}
+	s := bufio.NewScanner(r)
+	var b []byte
+	take := false
+	for s.Scan() {
+		line := s.Text()
+
+		if contains(strings.TrimSpace(line), dividers) {
+			take = !take
+			continue
+		}
+
+		if take {
+			b = append(b, []byte(line)...)
+		}
+	}
+	if err := s.Err(); err != nil {
+		return nil, err
+	}
+	return b, nil
+}
+
+func contains(needle string, haystack []string) bool {
+	for _, s := range haystack {
+		if s == needle {
+			return true
+		}
+	}
+	return false
+}
diff --git a/internal/asmdb/parse_test.go b/internal/asmdb/parse_test.go
new file mode 100644
index 00000000..e860c000
--- /dev/null
+++ b/internal/asmdb/parse_test.go
@@ -0,0 +1,13 @@
+package asmdb
+
+import "testing"
+
+func TestParseRawFile(t *testing.T) {
+	db, err := ParseRawFile("testdata/x86data.js")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(db.Instructions) == 0 {
+		t.FailNow()
+	}
+}
diff --git a/internal/asmdb/testdata b/internal/asmdb/testdata
new file mode 120000
index 00000000..4909e06e
--- /dev/null
+++ b/internal/asmdb/testdata
@@ -0,0 +1 @@
+../data
\ No newline at end of file