Compare commits

...

7 Commits

Author SHA1 Message Date
b2a588e667 Add "len" methods to encode and decode
Some checks failed
Benchmark BufferPool / RunBenchmarks (push) Failing after 18s
Run Tests / Test (push) Failing after 16s
To get length of encoded/decoded string
2024-07-30 22:45:56 +02:00
db61da9fc2 A lil optimize 2024-07-30 22:15:36 +02:00
9fbcdd86cc Add encode benchmarks 2024-07-30 22:14:19 +02:00
76d7102e72 Implement basic decoding 2024-07-30 22:10:57 +02:00
74928d9f7a Add some sort of work on decode 2024-07-30 22:03:57 +02:00
a6339587d4 Add tests for insertat and decode 2024-07-30 22:03:57 +02:00
48d28def47 Add air config 2024-07-30 22:02:00 +02:00
5 changed files with 351 additions and 40 deletions

1
.gitignore vendored
View File

@@ -1 +1,2 @@
jstester/node_modules
encoding/tmp

51
encoding/.air.toml Normal file
View File

@@ -0,0 +1,51 @@
root = "."
testdata_dir = "testdata"
tmp_dir = "tmp"
[build]
args_bin = []
bin = "tmp\\main.exe"
cmd = "go test ."
delay = 1000
exclude_dir = ["assets", "tmp", "vendor", "testdata"]
exclude_file = []
exclude_regex = []
exclude_unchanged = false
follow_symlink = false
full_bin = ""
include_dir = []
include_ext = ["go", "tpl", "tmpl", "html"]
include_file = []
kill_delay = "0s"
log = "build-errors.log"
poll = false
poll_interval = 0
post_cmd = []
pre_cmd = []
rerun = false
rerun_delay = 100
send_interrupt = false
stop_on_error = false
[color]
app = ""
build = "yellow"
main = "magenta"
runner = "green"
watcher = "cyan"
[log]
main_only = false
time = false
[misc]
clean_on_exit = true
[proxy]
app_port = 0
enabled = false
proxy_port = 0
[screen]
clear_on_rebuild = true
keep_scroll = true

View File

@@ -3,16 +3,23 @@ package encoding
import (
"bytes"
"fmt"
"log"
)
type GSM7Coder struct{}
// Make sure buffer can fit EncodesInto bytes
// Otherwise Encode will allocate memory as it sees fit
// Which is fine but not optimal
// Preallocate the buffer with the size of EncodesInto bytes
func (c *GSM7Coder) Encode(s string, buf *bytes.Buffer) error {
// utf8 := *(*[]byte)(unsafe.Pointer(&s))
utf8 := []byte(s)
var offset int = 1
var bitshift byte = 0
var leap, shift bool
var (
offset int = 1
bitshift byte = 0
leap, shift bool
)
for index, septet := range utf8 {
if septet > 0b01111111 {
@@ -56,9 +63,97 @@ func (c *GSM7Coder) Encode(s string, buf *bytes.Buffer) error {
offset = 1
}
}
log.Println(buf.Cap(), buf.Len())
return nil
}
func (c *GSM7Coder) Decode(buf *bytes.Buffer) (string, error) {
return buf.String(), nil
gsm7 := buf.Bytes()
var (
offset int = (len(gsm7) / 8) + 1
bitshift byte = 0
leap, shift bool
)
outLength := DecodesInto(buf)
lengthDiff := outLength - len(gsm7)
gsm7 = append(gsm7, make([]byte, lengthDiff)...)
// We don't care about the last byte
// Unless it's the %8.......
// We'll deal with that later
for index := len(gsm7) - 2; index >= 0; index-- {
octet := gsm7[index]
bitshift = byte((index + 1) % 8)
if bitshift == 7 {
leap = true
}
// log.Printf("Index:%-3d Offset:%-3d Bitshift:%-3d CurrentByte:%08b (%-3d) OffsetByte(%-3d):%08b (%-3d) Leap:%5v", index, offset, bitshift, gsm7[index], gsm7[index], index+offset, gsm7[index+offset], gsm7[index+offset], leap)
mask := byte(255 << (8 - bitshift))
masked := (mask & octet) >> (8 - bitshift)
// log.Printf("Index:%-3d Offset:%-3d Bitshift:%-3d Mask:%08b Masked:%08b", index, offset, bitshift, mask, masked)
if leap {
InsertAt(&gsm7, index+offset, masked)
} else {
gsm7[index+offset] |= masked
}
// Remove last bitshift bits
gsm7[index] <<= bitshift
// Move the remaining bit once to the right to form septet instead of octet
gsm7[index] >>= 1
// log.Printf("Index:%-3d Offset:%-3d Bitshift:%-3d CurrentByte:%08b (%-3d) OffsetByte(%-3d):%08b (%-3d) Leap:%5v", index, offset, bitshift, gsm7[index], gsm7[index], index+offset, gsm7[index+offset], gsm7[index+offset], leap)
leap = false
continue
if !leap {
buf.WriteByte(gsm7[index+offset])
}
if index == len(gsm7)-1 && gsm7[index] > 0 {
buf.WriteByte(gsm7[index])
}
// log.Printf("Index:%-3d Offset:%-3d Bitshift:%-3d CurrentByte:%08b (%-3d) OffsetByte:%08b (%-3d) Leap:%5v", index, offset, bitshift, gsm7[index], gsm7[index], gsm7[index-offset], gsm7[index-offset], leap)
if bitshift >= 7 {
if leap {
// log.Printf("Shift at Index:%-3d Offset:%-3d Bitshift:%-3d", index, offset, bitshift)
leap = false
bitshift = 0
offset++
shift = true
continue
}
// log.Printf("Leap at Index:%-3d Offset:%-3d Bitshift:%-3d", index, offset, bitshift)
leap = true
bitshift = 6
}
if shift {
offset = 1
}
}
// log.Printf("Result: %+v", gsm7)
// for _, v := range gsm7 {
// log.Printf("%08b", v)
// }
return string(gsm7), nil
}
// Allocation free
// Which means data MUST have space for value
func InsertAt(data *[]byte, index int, value byte) {
copy((*data)[index+1:], (*data)[index:])
(*data)[index] = value
}
func EncodesInto(s *string) int {
slen := len(*s)
enclen := slen * 7 / 8
if slen%8 != 0 {
enclen++
}
return enclen
}
func DecodesInto(buf *bytes.Buffer) int {
blen := buf.Len()
declen := blen * 8 / 7
return declen
}

View File

@@ -93,54 +93,217 @@ func TestGSM7DecodeSimpleASCIIString(t *testing.T) {
}
}
func TestGSM7DecodeComplexASCIIString(t *testing.T) {
coder := &GSM7Coder{}
var buf bytes.Buffer
// func TestGSM7DecodeComplexASCIIString(t *testing.T) {
// coder := &GSM7Coder{}
// var buf bytes.Buffer
// input := []byte{0b11000111, 0b00110111, 0b10011011, 0b01011100, 0b01110110, 0b10000011, 0b11100100, 0b11100001, 0b11111100, 0b00011100, 0b00000100, 0b01100111, 0b10000111, 0b11110011, 0b00101100, 0b11010000, 0b00010000, 0b00011101, 0b10011110, 0b10100111, 0b11011101, 0b01100111, 0b10010000, 0b00111011, 0b01111101, 0b01000110, 0b11010011, 0b01000001, 0b11100001, 0b01111011, 0b00111000, 0b11101111, 0b00000010}
// expected := "Golden rays play, Chasing night away."
// buf.Write(input)
// actual, err := coder.Decode(&buf)
// if err != nil {
// t.Errorf("Expected no error, but got %v", err)
// }
// if actual != expected {
// t.Errorf("Expected '%v', but got '%v'", expected, actual)
// }
// }
// func TestGSM7DecodeComplex8nASCIIString(t *testing.T) {
// coder := &GSM7Coder{}
// var buf bytes.Buffer
// input := []byte{0b11000100, 0b11111010, 0b01111000, 0b00111101, 0b00000111, 0b10000101, 0b11100101, 0b01100101, 0b10010000, 0b10111001, 0b00111110, 0b01011110, 0b10100111, 0b11011101, 0b01100111, 0b11010000, 0b01011001, 0b01011110, 0b00001110, 0b11010011, 0b01011001, 0b00100000, 0b00111010, 0b10111010, 0b10011100, 0b00000111, 0b11000101, 0b11101011, 0b11100001, 0b11110001, 0b01111010, 0b11001110, 0b00000010, 0b00111101, 0b01000001, 0b11110001, 0b01111010, 0b01111000, 0b10111100, 0b00101110, 0b11001011, 0b11100111, 0b00101100, 0b11010000, 0b00011101, 0b00011101, 0b10100110, 0b10000011, 0b11101000, 0b11101000, 0b00110010, 0b11001000, 0b01011100, 0b00011111, 0b10101111, 0b01011101}
// expected := "Ducks are fucking great, they quacks, O quackers, what the fuck."
// buf.Write(input)
// actual, err := coder.Decode(&buf)
// if err != nil {
// t.Errorf("Expected no error, but got %v", err)
// }
// if actual != expected {
// t.Errorf("Expected '%v', but got '%v'", expected, actual)
// }
// }
// func TestGSM7DecodeEmptyString(t *testing.T) {
// coder := &GSM7Coder{}
// buf := bytes.NewBuffer([]byte{})
// expected := ""
// actual, err := coder.Decode(buf)
// if err != nil {
// t.Errorf("Expected no error, but got %v", err)
// }
// if actual != expected {
// t.Errorf("Expected '%v', but got '%v'", expected, actual)
// }
// }
// region insertat
func TestInsertAtBeginning(t *testing.T) {
data := []byte{2, 3, 4, 0}
InsertAt(&data, 0, 1)
expected := []byte{1, 2, 3, 4}
if !bytes.Equal(data, expected) {
t.Errorf("expected %v, got %v", expected, data)
}
}
func TestInsertInMiddle(t *testing.T) {
data := []byte{2, 3, 4, 0}
InsertAt(&data, 1, 5)
expected := []byte{2, 5, 3, 4}
if !bytes.Equal(data, expected) {
t.Errorf("expected %v, got %v", expected, data)
}
}
func TestInsertAtEnd(t *testing.T) {
data := []byte{1, 2, 3, 0}
InsertAt(&data, 3, 4)
expected := []byte{1, 2, 3, 4}
if !bytes.Equal(data, expected) {
t.Errorf("expected %v, got %v", expected, data)
}
}
func TestIndexOutOfBounds(t *testing.T) {
data := []byte{2, 3, 4}
defer func() {
if r := recover(); r == nil {
t.Errorf("The code did not panic")
}
}()
InsertAt(&data, 4, 5)
}
func TestNegativeIndex(t *testing.T) {
data := []byte{2, 3, 4}
defer func() {
if r := recover(); r == nil {
t.Errorf("The code did not panic")
}
}()
InsertAt(&data, -1, 1)
}
func TestMaintainsOrderAfterInsertion(t *testing.T) {
data := []byte{2, 3, 4, 0}
InsertAt(&data, 1, 1)
expected := []byte{2, 1, 3, 4}
if !bytes.Equal(data, expected) {
t.Errorf("expected %v, got %v", expected, data)
}
}
func TestMaintainsLength(t *testing.T) {
data := []byte{2, 3, 4, 0}
InsertAt(&data, 1, 1)
if len(data) != 4 {
t.Errorf("expected length 4, got %d", len(data))
}
}
func TestDeletesLastValue(t *testing.T) {
data := []byte{2, 3, 4, 5}
InsertAt(&data, 0, 1)
expected := []byte{1, 2, 3, 4}
if !bytes.Equal(data, expected) {
t.Errorf("expected %v, got %v", expected, data)
}
}
// region misc tests
func TestGSM7EncodesIntoSmallString(t *testing.T) {
input := "Sunshine"
expected := 7
actual := EncodesInto(&input)
if actual != expected {
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
func TestGSM7EncodesIntoLargerNot8nString(t *testing.T) {
input := "Golden rays play, Chasing night away."
expected := 33
actual := EncodesInto(&input)
if actual != expected {
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
func TestGSM7EncodesIntoLarger8nString(t *testing.T) {
input := "Ducks are fucking great, they quacks, O quackers, what the fuck."
expected := 56
actual := EncodesInto(&input)
if actual != expected {
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
func TestGSM7DecodesIntoSmallString(t *testing.T) {
input := []byte{0b11010011, 0b10111010, 0b01111011, 0b10001110, 0b01001110, 0b10111011, 0b11001011}
expected := "Golden rays play, Chasing night away."
buf.Write(input)
actual, err := coder.Decode(&buf)
if err != nil {
t.Errorf("Expected no error, but got %v", err)
}
expected := 8
actual := DecodesInto(bytes.NewBuffer(input))
if actual != expected {
t.Errorf("Expected '%v', but got '%v'", expected, actual)
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
func TestGSM7DecodeComplex8nASCIIString(t *testing.T) {
func TestGSM7DecodesIntoLargerNot8nString(t *testing.T) {
input := []byte{0b11000111, 0b00110111, 0b10011011, 0b01011100, 0b01110110, 0b10000011, 0b11100100, 0b11100001, 0b11111100, 0b00011100, 0b00000100, 0b01100111, 0b10000111, 0b11110011, 0b00101100, 0b11010000, 0b00010000, 0b00011101, 0b10011110, 0b10100111, 0b11011101, 0b01100111, 0b10010000, 0b00111011, 0b01111101, 0b01000110, 0b11010011, 0b01000001, 0b11100001, 0b01111011, 0b00111000, 0b11101111, 0b00000010}
expected := 37
actual := DecodesInto(bytes.NewBuffer(input))
if actual != expected {
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
func TestGSM7DecodesIntoLarger8nString(t *testing.T) {
input := []byte{0b11000100, 0b11111010, 0b01111000, 0b00111101, 0b00000111, 0b10000101, 0b11100101, 0b01100101, 0b10010000, 0b10111001, 0b00111110, 0b01011110, 0b10100111, 0b11011101, 0b01100111, 0b11010000, 0b01011001, 0b01011110, 0b00001110, 0b11010011, 0b01011001, 0b00100000, 0b00111010, 0b10111010, 0b10011100, 0b00000111, 0b11000101, 0b11101011, 0b11100001, 0b11110001, 0b01111010, 0b11001110, 0b00000010, 0b00111101, 0b01000001, 0b11110001, 0b01111010, 0b01111000, 0b10111100, 0b00101110, 0b11001011, 0b11100111, 0b00101100, 0b11010000, 0b00011101, 0b00011101, 0b10100110, 0b10000011, 0b11101000, 0b11101000, 0b00110010, 0b11001000, 0b01011100, 0b00011111, 0b10101111, 0b01011101}
expected := 64
actual := DecodesInto(bytes.NewBuffer(input))
if actual != expected {
t.Errorf("Expected %d, but got %d", expected, actual)
}
}
// region benchmark
func BenchmarkGSM7EncodeSimpleASCIIString(b *testing.B) {
coder := &GSM7Coder{}
var buf bytes.Buffer
input := []byte{0b11000100, 0b11111010, 0b01111000, 0b00111101, 0b00000111, 0b10000101, 0b11100101, 0b01100101, 0b10010000, 0b10111001, 0b00111110, 0b01011110, 0b10100111, 0b11011101, 0b01100111, 0b11010000, 0b01011001, 0b01011110, 0b00001110, 0b11010011, 0b01011001, 0b00100000, 0b00111010, 0b10111010, 0b10011100, 0b00000111, 0b11000101, 0b11101011, 0b11100001, 0b11110001, 0b01111010, 0b11001110, 0b00000010, 0b00111101, 0b01000001, 0b11110001, 0b01111010, 0b01111000, 0b10111100, 0b00101110, 0b11001011, 0b11100111, 0b00101100, 0b11010000, 0b00011101, 0b00011101, 0b10100110, 0b10000011, 0b11101000, 0b11101000, 0b00110010, 0b11001000, 0b01011100, 0b00011111, 0b10101111, 0b01011101}
input := "Sunshine"
b.ResetTimer()
expected := "Ducks are fucking great, they quacks, O quackers, what the fuck."
buf.Write(input)
actual, err := coder.Decode(&buf)
if err != nil {
t.Errorf("Expected no error, but got %v", err)
}
if actual != expected {
t.Errorf("Expected '%v', but got '%v'", expected, actual)
for i := 0; i < b.N; i++ {
coder.Encode(input, &buf)
}
}
func TestGSM7DecodeEmptyString(t *testing.T) {
func BenchmarkGSM7EncodeComplexASCIIString(b *testing.B) {
coder := &GSM7Coder{}
buf := bytes.NewBuffer([]byte{})
var buf bytes.Buffer
input := "Golden rays play, Chasing night away."
b.ResetTimer()
expected := ""
actual, err := coder.Decode(buf)
if err != nil {
t.Errorf("Expected no error, but got %v", err)
}
if actual != expected {
t.Errorf("Expected '%v', but got '%v'", expected, actual)
for i := 0; i < b.N; i++ {
coder.Encode(input, &buf)
}
}
func BenchmarkGSM7EncodeComplex8nASCIIString(b *testing.B) {
coder := &GSM7Coder{}
var buf bytes.Buffer
input := "Ducks are fucking great, they quacks, O quackers, what the fuck."
b.ResetTimer()
for i := 0; i < b.N; i++ {
coder.Encode(input, &buf)
}
}

View File

@@ -78,6 +78,7 @@ func (p *SUBMIT_SM) Encode(buf *bytes.Buffer) error {
// UCS2 should also be fairly easy, use uint16 or something
// GSM7 will not be easy
// See https://www.codeproject.com/Tips/470755/Encoding-Decoding-7-bit-User-Data-for-SMS-PDU-PDU
// Another great site: https://doubleblak.com/blogPost.php?k=7bitpdu
asciiEncoder := ascii85.NewEncoder(buf)
// TODO: Implement encodings bsed on p.data_coding
messageEncoder := gsm7.NewEncoder()