hacked parity context version for arm code

Cargo.lock

@@ -62,12 +62,6 @@ version = "1.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
-[[package]]
-name = "lexopt"
-version = "0.2.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "478ee9e62aaeaf5b140bd4138753d1f109765488581444218d3ddda43234f3e8"
-
 [[package]]
 name = "libc"
 version = "0.2.108"
@@ -95,6 +89,12 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "pico-args"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "db8bcd96cb740d03149cbad5518db9fd87126a10ab519c011893b1754134c468"
+
 [[package]]
 name = "sacabase"
 version = "2.0.0"
@@ -121,8 +121,8 @@ version = "0.1.0"
 dependencies = [
  "anyhow",
  "cdivsufsort",
- "lexopt",
  "pbr",
+ "pico-args",
 ]
 
 [[package]]

Cargo.toml

@@ -1,12 +1,12 @@
 [package]
 name = "upkr"
-version = "0.2.0"
+version = "0.1.0"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
 cdivsufsort = "2"
-lexopt = "0.2.1"
+pico-args = "0.4"
 anyhow = "1"
 pbr = "1"

asm_unpackers/.gitignore

@@ -1 +0,0 @@
-/build/

asm_unpackers/Makefile

@@ -1,36 +0,0 @@
-build/unpack_riscv64: ../c_unpacker/main.c unpack_riscv.S
-	mkdir -p build
-	riscv64-linux-gnu-gcc -g -static -o $@ $^
-
-test_riscv64: build/unpack_riscv64
-	qemu-riscv64 $< test_data.upk /tmp/out.bin
-	cmp test_data.bin /tmp/out.bin
-
-build/unpack_riscv64.bin: unpack_riscv.S
-	mkdir -p build
-	riscv64-linux-gnu-gcc -c -o build/unpack_riscv64.o $?
-	riscv64-linux-gnu-objcopy -O binary --only-section=.text build/unpack_riscv64.o $@
-
-build/unpack_armv6m: ../c_unpacker/main.c unpack_armv6m.S
-	mkdir -p build
-	arm-linux-gnueabihf-gcc -g -static -o $@ $^
-
-test_armv6m: build/unpack_armv6m
-	qemu-arm $< test_data.upk /tmp/out.bin
-	cmp test_data.bin /tmp/out.bin
-
-build/unpack_armv6m.bin: unpack_armv6m.S
-	mkdir -p build
-	arm-none-eabi-gcc -march=armv6-m -c -o build/unpack_armv6m.o $?
-	arm-none-eabi-objcopy -O binary --only-section=.text build/unpack_armv6m.o $@
-	
-build/unpack_c: ../c_unpacker/main.c ../c_unpacker/unpack.c
-	mkdir -p build
-	gcc -g -o $@ $^
-
-test_c: build/unpack_c
-	$< test_data.upk /tmp/out.bin
-	cmp test_data.bin /tmp/out.bin
-
-sizes: build/unpack_armv6m.bin build/unpack_riscv64.bin
-	ls -l build/*.bin

asm_unpackers/test_data.bin

@@ -1,99 +0,0 @@
-typedef unsigned char u8;
-typedef unsigned short u16;
-typedef unsigned long u32;
-
-u8* upkr_data_ptr;
-u8 upkr_probs[1 + 255 + 1 + 2*32 + 2*32]; 
-#ifdef UPKR_BITSTREAM
-u16 upkr_state;
-u8 upkr_current_byte;
-int upkr_bits_left;
-#else
-u32 upkr_state;
-#endif
-
-int upkr_decode_bit(int context_index) {
-#ifdef UPKR_BITSTREAM
-    while(upkr_state < 32768) {
-        if(upkr_bits_left == 0) {
-            upkr_current_byte = *upkr_data_ptr++;
-            upkr_bits_left = 8;
-        }
-        upkr_state = (upkr_state << 1) + (upkr_current_byte & 1);
-        upkr_current_byte >>= 1;
-        --upkr_bits_left;
-    }
-#else
-    while(upkr_state < 4096) {
-        upkr_state = (upkr_state << 8) | *upkr_data_ptr++;
-    }
-#endif
-   
-    int prob = upkr_probs[context_index];
-    int bit = (upkr_state & 255) < prob ? 1 : 0;
-    
-    int tmp = prob;
-    if(!bit) {
-        tmp = 256 - tmp;
-    }
-    upkr_state = tmp * (upkr_state >> 8) + (upkr_state & 255);
-    tmp += (256 - tmp + 8) >> 4;
-    if(!bit) {
-        upkr_state -= prob;
-        tmp = 256 - tmp;
-    }
-    upkr_probs[context_index] = tmp;
-
-    return bit;
-}
-
-int upkr_decode_length(int context_index) {
-    int length = 0;
-    int bit_pos = 0;
-    while(upkr_decode_bit(context_index)) {
-        length |= upkr_decode_bit(context_index + 1) << bit_pos++;
-        context_index += 2;
-    }
-    return length | (1 << bit_pos);
-}
-
-void* upkr_unpack(void* destination, void* compressed_data) {
-    upkr_data_ptr = (u8*)compressed_data;
-    upkr_state = 0;
-#ifdef UPKR_BITSTREAM
-    upkr_bits_left = 0;
-#endif
-    for(int i = 0; i < sizeof(upkr_probs); ++i)
-        upkr_probs[i] = 128;
-    
-    u8* write_ptr = (u8*)destination;
-    
-    int prev_was_match = 0;
-    int offset = 0;
-    for(;;) {
-        if(upkr_decode_bit(0)) {
-            if(prev_was_match || upkr_decode_bit(256)) {
-                offset = upkr_decode_length(257) - 1;
-                if(offset == 0) {
-                    break;
-                }
-            }
-            int length = upkr_decode_length(257 + 64);
-            while(length--) {
-                *write_ptr = write_ptr[-offset];
-                ++write_ptr;
-            }
-            prev_was_match = 1;
-        } else {
-            int byte = 1;
-            while(byte < 256) {
-                int bit = upkr_decode_bit(byte);
-                byte = (byte << 1) + bit;
-            }
-            *write_ptr++ = byte;
-            prev_was_match = 0;
-        }
-    }
-    
-    return write_ptr;
-}

asm_unpackers/unpack_armv6m.S

@@ -1,162 +0,0 @@
-// armv6-m upkr unpacker by yrlf
-// some optimizations by exoticorn
-
-.syntax unified
-.thumb
-
-.section .text
-
-#define ALIGNUP(n, align) (((n) + (align) - 1) & ~((align) - 1))
-#define PROB_LEN (1 + 255 + 1 + 2*32 + 2*32)
-#define FRAME_SIZE ALIGNUP(PROB_LEN, 4)
-
-// auto upkr_unpack(uint8_t * out, uint8_t * in) -> tuple<uint8_t *, uint8_t *>
-.global upkr_unpack
-.type upkr_unpack, %function
-// r0 .. out_ptr (returned)
-// r1 .. in_ptr (returned)
-// r2 .. state
-// r3 .. offset
-// r4 .. prev_was_literal / decode_length ret
-// r5 .. subroutine arg (preserved)
-// r6 .. decode_bit ret
-// r7 .. probs ptr
-upkr_unpack:
-    push { r4, r5, r6, r7, lr }
-    sub sp, sp, #FRAME_SIZE
-
-    mov r7, sp
-    movs r2, #255
-    adds r2, r2, #(PROB_LEN - 255)
-    movs r3, #128
-.Lclear:
-    subs r2, r2, #1
-    strb r3, [r7, r2]
-    bne .Lclear
-
-.Lloop:
-    movs r5, #0
-    bl upkr_decode_bit
-    beq .Ldata
-.Lmatch:
-    // r6 = 1
-    lsls r5, r6, #8
-    cmp r4, #0
-    beq 1f
-
-    bl upkr_decode_bit
-    beq 2f
-
-1:
-    bl upkr_decode_length
-    adds r3, r4, #1
-    beq .Lend
-2:
-
-    adds r5, r5, #64
-    bl upkr_decode_length
-.Lcopy_loop:
-    ldrb r5, [r0, r3]
-.Lstore:
-    strb r5, [r0]
-    adds r0, r0, #1
-    adds r4, r4, #1
-    blt .Lcopy_loop
-    b .Lloop
-
-.Ldata:
-    movs r5, #1
-
-.Ldata_loop:
-    bl upkr_decode_bit
-    adcs r5, r5, r5
-    lsrs r4, r5, #8
-    beq .Ldata_loop
-    b .Lstore
-
-.Lend:
-    add sp, sp, #FRAME_SIZE
-    pop { r4, r5, r6, r7, pc }
-
-.type upkr_decode_length, %function
-// r0 .. -length tmp (saved)
-// r1 ..
-// r2 ..
-// r3 ..
-// r4 .. -length (returned)
-// r5 .. context index (saved)
-// r6 .. (saved)
-// r7 ..
-upkr_decode_length:
-    push { r0, r5, r6, lr }
-
-    movs r0, #0
-    subs r4, r0, #1
-.Lbit_loop:
-    adds r5, r5, #1
-    bl upkr_decode_bit
-    beq 1f
-
-    adds r5, r5, #1
-    bl upkr_decode_bit
-    beq 2f
-    adds r0, r0, r4
-2:
-    lsls r4, r4, #1
-    b .Lbit_loop
-1:
-    adds r4, r4, r0
-
-    pop { r0, r5, r6, pc }
-
-.type upkr_decode_bit, %function
-// r0 .. tmp / prob (saved)
-// r1 .. in_ptr (modified)
-// r2 .. state (modified)
-// r3 .. scratch (saved)
-// r4 ..
-// r5 .. context index (preserved)
-// r6 .. bit (returned)
-// r7 .. probs ptr (preserved)
-upkr_fill_state:
-    lsls r2, r2, #8
-    ldrb r6, [r1]
-    adds r1, r1, #1
-    orrs r2, r2, r6
-
-upkr_decode_bit:
-    lsrs r6, r2, #12
-    beq upkr_fill_state
-
-    push { r0, r1, r3, lr }
-
-    ldrb r0, [r7, r5]
-
-    lsrs r3, r2, #8
-    uxtb r1, r2
-
-    subs r6, r1, r0
-    blt 1f
-
-    subs r1, r2, r0
-    rsbs r0, r0, #0
-1:
-
-    muls r3, r3, r0
-    adds r2, r1, r3
-
-    rsbs r3, r0, #0
-    uxtb r3, r3
-    lsrs r3, r3, #4
-    adcs r0, r0, r3
-
-    cmp r6, #0
-    blt 1f
-
-    rsbs r0, r0, #0
-1:
-
-    strb r0, [r7, r5]
-
-    lsrs r6, r6, #31
-    pop { r0, r1, r3, pc }

asm_unpackers/unpack_riscv.S

@@ -1,144 +0,0 @@
-.section .text
-
-#define FRAME_SIZE (256+64*4+4)
-
-// x8 prob array ptr
-// x9 prev was literal
-// x10 out ptr
-// x11 in ptr
-// x12 offset
-// x13 state
-
-.global upkr_unpack
-.type upkr_unpack, %function
-upkr_unpack:
-	mv t4, ra
-	mv x17, x8
-	mv t6, x9
-	li x13, FRAME_SIZE
-	li x9, 128
-1:
-	addi sp, sp, -1
-	sb x9, 0(sp)
-	addi x13, x13, -1
-	bnez x13, 1b
-	
-.Lmainloop:
-	li x14, 0
-	jal upkr_decode_bit
-	beqz x15, .Lliteral
-	
-	li x14, 256
-	beqz x9, .Lread_offset
-	jal upkr_decode_bit
-	beqz x15, .Lskip_offset
-.Lread_offset:
-	jal t3, upkr_decode_number
-	addi x12, x9, -1
-	beqz x12, .Ldone
-
-.Lskip_offset:
-	li x14, 256+64
-	jal t3, upkr_decode_number
-1:
-	sub x15, x10, x12
-	lbu x15, (x15)
-	sb x15, (x10)
-	addi x10, x10, 1
-	addi x9, x9, -1
-	bnez x9, 1b
-	j .Lmainloop
-	
-.Lliteral:
-	li x14, 1
-1:
-	jal upkr_decode_bit
-	slli x14, x14, 1
-	add x14, x14, x15
-	srli x9, x14, 8
-	beqz x9, 1b
-	sb x14, 0(x10)
-	addi x10, x10, 1
-	j .Lmainloop
-
-.Ldone:
-	addi sp, sp, FRAME_SIZE
-	mv x8, x17
-	mv x9, t6
-	jr t4
-
-// x14 context index
-// return: x9 decoded number 
-upkr_decode_number:
-	mv t5, x14
-	li x9, 0
-	li x8, 1
-1:
-	addi x14, x14, 1
-	jal upkr_decode_bit
-	beqz x15, 1f
-	addi x14, x14, 1
-	jal upkr_decode_bit
-	beqz x15, 2f
-	add x9, x9, x8
-2:
-	slli x8, x8, 1
-	j 1b
-1:
-	add x9, x9, x8
-	
-	mv x14, t5
-	jr t3
-
-upkr_load_byte:
-	lbu x15, 0(x11)
-	addi x11, x11, 1
-	slli x13, x13, 8
-	add x13, x13, x15
-// x8 prob array ptr
-// x11 in ptr
-// x13 state
-// x14 context index
-// return: x15 decoded bit
-upkr_decode_bit:
-	srli x15, x13, 12
-	beqz x15, upkr_load_byte
-
-	mv t0, x9
-	mv t1, x14
-	mv t2, x10
-	
-	add x14, x14, sp
-	lbu x9, 0(x14)
-	
-	andi x10, x13, 255
-	sltu x15, x10, x9
-	srli x13, x13, 8
-	beqz x15, .Lelse
-	
-	mul x13, x13, x9
-	add x13, x13, x10
-	li x10, 256 + 8
-	sub x10, x10, x9
-	srli x10, x10, 4
-	add x9, x9, x10
-	j .Lendif
-
-.Lelse:
-	li x16, 256
-	sub x16, x16, x9
-	mul x13, x13, x16
-	add x13, x13, x10
-	sub x13, x13, x9
-	addi x10, x9, 8
-	srli x10, x10, 4
-	sub x9, x9, x10
-
-.Lendif:
-
-	sb x9, 0(x14)
-	
-	mv x9, t0
-	mv x14, t1
-	mv x10, t2
-	ret

c_unpacker/.gitignore

@@ -1,5 +0,0 @@
-unpack
-unpack_bitstream
-unpack_debug
-*.upk
-

c_unpacker/Makefile

@@ -1,10 +0,0 @@
-all: unpack unpack_bitstream
-
-unpack: main.c unpack.c
-	cc -O2 -o unpack main.c unpack.c
-
-unpack_bitstream: main.c unpack.c
-	cc -O2 -D UPKR_BITSTREAM -o unpack_bitstream main.c unpack.c
-	
-unpack_debug: main.c unpack.c
-	cc -g -o unpack_debug main.c unpack.c

c_unpacker/decode_bit_alt.c

@@ -1,33 +0,0 @@
-int upkr_decode_bit(int context_index) {
-#ifdef UPKR_BITSTREAM
-    while(upkr_state < 32768) {
-        if(upkr_bits_left == 0) {
-            upkr_current_byte = *upkr_data_ptr++;
-            upkr_bits_left = 8;
-        }
-        upkr_state = (upkr_state << 1) + (upkr_current_byte & 1);
-        upkr_current_byte >>= 1;
-        --upkr_bits_left;
-    }
-#else
-    while(upkr_state < 4096) {
-        upkr_state = (upkr_state << 8) | *upkr_data_ptr++;
-    }
-#endif
-   
-    int prob = upkr_probs[context_index];
-    int bit = (upkr_state & 255) < prob ? 1 : 0;
-    
-    if(bit) {
-        prob = 256 - prob;
-    }
-    upkr_state -= prob * ((upkr_state >> 8) + (bit ^ 1));
-    prob -= (prob + 8) >> 4;
-    if(bit) {
-        prob = -prob;
-    }
-    upkr_probs[context_index] = prob;
-
-    return bit;
-}
-

c_unpacker/main.c

@@ -1,26 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-
-void* upkr_unpack(void* destination, void* compressed_data);
-
-int main(int argn, char** argv) {
-  void* input_buffer = malloc(1024*1024);
-  void* output_buffer = malloc(1024*1024);
-  
-  FILE* in_file = fopen(argv[1], "rb");
-  int in_size = fread(input_buffer, 1, 1024*1024, in_file);
-  fclose(in_file);
-  
-  printf("Compressed size: %d\n", in_size);
-  
-  void* end_ptr = upkr_unpack(output_buffer, input_buffer);
-  int out_size = (char*)end_ptr - (char*)output_buffer;
-  
-  printf("Uncompressed size: %d\n", out_size);
-  
-  FILE* out_file = fopen(argv[2], "wb");
-  fwrite(output_buffer, 1, out_size, out_file);
-  fclose(out_file);
-  
-  return 0;
-}

c_unpacker/readme.txt

@@ -1,4 +0,0 @@
-a very simple unpacker in c, as a reference for people wanting to implement their own unpacker.
-absolutely not production ready, it makes no effort to ensure the output buffer can actually
-hold the uncompressed data.
-!!! Never run on untrusted input !!!

c_unpacker/unpack.c

@@ -1,96 +0,0 @@
-typedef unsigned char u8;
-typedef unsigned short u16;
-typedef unsigned long u32;
-
-u8* upkr_data_ptr;
-u8 upkr_probs[1 + 255 + 1 + 2*32 + 2*32]; 
-#ifdef UPKR_BITSTREAM
-u16 upkr_state;
-u8 upkr_current_byte;
-int upkr_bits_left;
-#else
-u32 upkr_state;
-#endif
-
-int upkr_decode_bit(int context_index) {
-#ifdef UPKR_BITSTREAM
-    while(upkr_state < 32768) {
-        if(upkr_bits_left == 0) {
-            upkr_current_byte = *upkr_data_ptr++;
-            upkr_bits_left = 8;
-        }
-        upkr_state = (upkr_state << 1) + (upkr_current_byte & 1);
-        upkr_current_byte >>= 1;
-        --upkr_bits_left;
-    }
-#else
-    while(upkr_state < 4096) {
-        upkr_state = (upkr_state << 8) | *upkr_data_ptr++;
-    }
-#endif
-   
-    int prob = upkr_probs[context_index];
-    int bit = (upkr_state & 255) < prob ? 1 : 0;
-    
-    if(bit) {
-        upkr_state = prob * (upkr_state >> 8) + (upkr_state & 255);
-        prob += (256 - prob + 8) >> 4;
-    } else {
-        upkr_state = (256 - prob) * (upkr_state >> 8) + (upkr_state & 255) - prob;
-        prob -= (prob + 8) >> 4;
-    }
-    upkr_probs[context_index] = prob;
-
-    return bit;
-}
-
-int upkr_decode_length(int context_index) {
-    int length = 0;
-    int bit_pos = 0;
-    while(upkr_decode_bit(context_index)) {
-        length |= upkr_decode_bit(context_index + 1) << bit_pos++;
-        context_index += 2;
-    }
-    return length | (1 << bit_pos);
-}
-
-void* upkr_unpack(void* destination, void* compressed_data) {
-    upkr_data_ptr = (u8*)compressed_data;
-    upkr_state = 0;
-#ifdef UPKR_BITSTREAM
-    upkr_bits_left = 0;
-#endif
-    for(int i = 0; i < sizeof(upkr_probs); ++i)
-        upkr_probs[i] = 128;
-    
-    u8* write_ptr = (u8*)destination;
-    
-    int prev_was_match = 0;
-    int offset = 0;
-    for(;;) {
-        if(upkr_decode_bit(0)) {
-            if(prev_was_match || upkr_decode_bit(256)) {
-                offset = upkr_decode_length(257) - 1;
-                if(offset == 0) {
-                    break;
-                }
-            }
-            int length = upkr_decode_length(257 + 64);
-            while(length--) {
-                *write_ptr = write_ptr[-offset];
-                ++write_ptr;
-            }
-            prev_was_match = 1;
-        } else {
-            int byte = 1;
-            while(byte < 256) {
-                int bit = upkr_decode_bit(byte);
-                byte = (byte << 1) + bit;
-            }
-            *write_ptr++ = byte;
-            prev_was_match = 0;
-        }
-    }
-    
-    return write_ptr;
-}

src/greedy_packer.rs

@@ -6,12 +6,11 @@ use crate::ProgressCallback;
 pub fn pack(
     data: &[u8],
     use_bitstream: bool,
-    parity_contexts: usize,
     mut progress_callback: Option<ProgressCallback>,
 ) -> Vec<u8> {
     let mut match_finder = MatchFinder::new(data);
     let mut rans_coder = RansCoder::new(use_bitstream);
-    let mut state = lz::CoderState::new(parity_contexts);
+    let mut state = lz::CoderState::new();
 
     let mut pos = 0;
     while pos < data.len() {

src/lib.rs

@@ -9,49 +9,15 @@ pub use lz::unpack;
 
 pub type ProgressCallback<'a> = &'a mut dyn FnMut(usize);
 
-pub struct Config {
-    pub use_bitstream: bool,
-    pub parity_contexts: usize,
-}
-
-impl Default for Config {
-    fn default() -> Config {
-        Config {
-            use_bitstream: false,
-            parity_contexts: 1,
-        }
-    }
-}
-
 pub fn pack(
     data: &[u8],
     level: u8,
-    config: Config,
+    use_bitstream: bool,
     progress_callback: Option<ProgressCallback>,
 ) -> Vec<u8> {
     if level == 0 {
-        greedy_packer::pack(
-            data,
-            config.use_bitstream,
-            config.parity_contexts,
-            progress_callback,
-        )
+        greedy_packer::pack(data, use_bitstream, progress_callback)
     } else {
-        parsing_packer::pack(
-            data,
-            level,
-            config.use_bitstream,
-            config.parity_contexts,
-            progress_callback,
-        )
+        parsing_packer::pack(data, level, use_bitstream, progress_callback)
     }
 }
-
-pub fn compressed_size(mut data: &[u8]) -> f32 {
-    let mut state = 0;
-    while state < 4096 {
-        state = (state << 8) | data[0] as u32;
-        data = &data[1..];
-    }
-    data.len() as f32 + (state as f32).log2() / 8.
-}

src/lz.rs

@@ -9,49 +9,44 @@ pub enum Op {
 
 impl Op {
     pub fn encode(&self, coder: &mut dyn EntropyCoder, state: &mut CoderState) {
-        let literal_base = state.pos % state.parity_contexts * 256;
+        let base_context = 256 * (state.pos & 3);
         match self {
             &Op::Literal(lit) => {
-                encode_bit(coder, state, literal_base, false);
+                encode_bit(coder, state, base_context, false);
                 let mut context_index = 1;
                 for i in (0..8).rev() {
                     let bit = (lit >> i) & 1 != 0;
-                    encode_bit(coder, state, literal_base + context_index, bit);
+                    encode_bit(coder, state, base_context + context_index, bit);
                     context_index = (context_index << 1) | bit as usize;
                 }
-                state.prev_was_match = false;
                 state.pos += 1;
+                state.prev_was_match = false;
             }
             &Op::Match { offset, len } => {
-                encode_bit(coder, state, literal_base, true);
+                encode_bit(coder, state, base_context, true);
                 if !state.prev_was_match {
-                    encode_bit(
-                        coder,
-                        state,
-                        256 * state.parity_contexts,
-                        offset != state.last_offset,
-                    );
+                    encode_bit(coder, state, 1024, offset != state.last_offset);
                 } else {
                     assert!(offset != state.last_offset);
                 }
                 if offset != state.last_offset {
-                    encode_length(coder, state, 256 * state.parity_contexts + 1, offset + 1);
+                    encode_length(coder, state, 1025, offset + 1);
                     state.last_offset = offset;
                 }
-                encode_length(coder, state, 256 * state.parity_contexts + 65, len);
-                state.prev_was_match = true;
+                encode_length(coder, state, 1025 + 64, len);
                 state.pos += len as usize;
+                state.prev_was_match = true;
             }
         }
     }
 }
 
 pub fn encode_eof(coder: &mut dyn EntropyCoder, state: &mut CoderState) {
-    encode_bit(coder, state, state.pos % state.parity_contexts * 256, true);
+    encode_bit(coder, state, 256 * (state.pos & 3), true);
     if !state.prev_was_match {
-        encode_bit(coder, state, 256 * state.parity_contexts, true);
+        encode_bit(coder, state, 1024, true);
     }
-    encode_length(coder, state, 256 * state.parity_contexts + 1, 1);
+    encode_length(coder, state, 1025, 1);
 }
 
 fn encode_bit(
@@ -84,20 +79,18 @@ fn encode_length(
 #[derive(Clone)]
 pub struct CoderState {
     contexts: ContextState,
-    parity_contexts: usize,
     last_offset: u32,
-    prev_was_match: bool,
     pos: usize,
+    prev_was_match: bool,
 }
 
 impl CoderState {
-    pub fn new(parity_contexts: usize) -> CoderState {
+    pub fn new() -> CoderState {
         CoderState {
-            contexts: ContextState::new((1 + 255) * parity_contexts + 1 + 64 + 64),
+            contexts: ContextState::new((1 + 255) * 4 + 1 + 64 + 64),
             last_offset: 0,
-            parity_contexts,
-            prev_was_match: false,
             pos: 0,
+            prev_was_match: false,
         }
     }
 
@@ -106,9 +99,9 @@ impl CoderState {
     }
 }
 
-pub fn unpack(packed_data: &[u8], config: crate::Config) -> Vec<u8> {
-    let mut decoder = RansDecoder::new(packed_data, config.use_bitstream);
-    let mut contexts = ContextState::new((1 + 255) * config.parity_contexts + 1 + 64 + 64);
+pub fn unpack(packed_data: &[u8], use_bitstream: bool) -> Vec<u8> {
+    let mut decoder = RansDecoder::new(packed_data, use_bitstream);
+    let mut contexts = ContextState::new((1 + 255) * 4 + 1 + 64 + 64);
     let mut result = vec![];
     let mut offset = 0;
     let mut prev_was_match = false;
@@ -131,26 +124,15 @@ pub fn unpack(packed_data: &[u8], config: crate::Config) -> Vec<u8> {
     }
 
     loop {
-        let literal_base = result.len() % config.parity_contexts * 256;
-        if decoder.decode_with_context(&mut contexts.context_mut(literal_base)) {
-            if prev_was_match
-                || decoder
-                    .decode_with_context(&mut contexts.context_mut(256 * config.parity_contexts))
-            {
-                offset = decode_length(
-                    &mut decoder,
-                    &mut contexts,
-                    256 * config.parity_contexts + 1,
-                ) - 1;
+        let base_context = 256 * (result.len() & 3);
+        if decoder.decode_with_context(&mut contexts.context_mut(base_context)) {
+            if prev_was_match || decoder.decode_with_context(&mut contexts.context_mut(1024)) {
+                offset = decode_length(&mut decoder, &mut contexts, 1025) - 1;
                 if offset == 0 {
                     break;
                 }
             }
-            let length = decode_length(
-                &mut decoder,
-                &mut contexts,
-                256 * config.parity_contexts + 65,
-            );
+            let length = decode_length(&mut decoder, &mut contexts, 1025 + 64);
             for _ in 0..length {
                 result.push(result[result.len() - offset]);
             }
@@ -159,8 +141,7 @@ pub fn unpack(packed_data: &[u8], config: crate::Config) -> Vec<u8> {
             let mut context_index = 1;
             let mut byte = 0;
             for i in (0..8).rev() {
-                let bit = decoder
-                    .decode_with_context(&mut contexts.context_mut(literal_base + context_index));
+                let bit = decoder.decode_with_context(&mut contexts.context_mut(base_context + context_index));
                 context_index = (context_index << 1) | bit as usize;
                 byte |= (bit as u8) << i;
             }

src/main.rs

@@ -1,115 +1,64 @@
-use anyhow::Result;
-use std::ffi::OsStr;
+use anyhow::{bail, Result};
 use std::io::prelude::*;
-use std::process;
 use std::{fs::File, path::PathBuf};
 
 fn main() -> Result<()> {
-    let mut config = upkr::Config::default();
-    let mut reverse = false;
-    let mut unpack = false;
-    let mut level = 2;
-    let mut infile: Option<PathBuf> = None;
-    let mut outfile: Option<PathBuf> = None;
+    let mut args = pico_args::Arguments::from_env();
 
-    let mut parser = lexopt::Parser::from_env();
-    while let Some(arg) = parser.next()? {
-        use lexopt::prelude::*;
-        match arg {
-            Short('b') | Long("bitstream") => config.use_bitstream = true,
-            Short('p') | Long("parity") => config.parity_contexts = parser.value()?.parse()?,
-            Short('r') | Long("reverse") => reverse = true,
-            Short('u') | Long("unpack") => unpack = true,
-            Short('l') | Long("level") => level = parser.value()?.parse()?,
-            Short('h') | Long("help") => print_help(0),
-            Value(val) if infile.is_none() => infile = Some(val.try_into()?),
-            Value(val) if outfile.is_none() => outfile = Some(val.try_into()?),
-            _ => return Err(arg.unexpected().into()),
+    match args.subcommand()?.as_ref().map(|s| s.as_str()) {
+        None => print_help(),
+        Some("pack") => {
+            let level = args.opt_value_from_str(["-l", "--level"])?.unwrap_or(2u8);
+            let use_bitstream = args.contains(["-b", "--bitstream"]);
+
+            let infile = args.free_from_os_str::<PathBuf, bool>(|s| Ok(s.into()))?;
+            let outfile = args.free_from_os_str::<PathBuf, bool>(|s| Ok(s.into()))?;
+
+            let mut data = vec![];
+            File::open(infile)?.read_to_end(&mut data)?;
+            
+            let mut pb = pbr::ProgressBar::new(data.len() as u64);
+            pb.set_units(pbr::Units::Bytes);
+            let packed_data = upkr::pack(
+                &data,
+                level,
+                use_bitstream,
+                Some(&mut |pos| {
+                    pb.set(pos as u64);
+                }),
+            );
+            pb.finish();
+
+            println!(
+                "Compressed {} bytes to {} bytes ({}%)",
+                data.len(),
+                packed_data.len(),
+                packed_data.len() as f32 * 100. / data.len() as f32
+            );
+            File::create(outfile)?.write_all(&packed_data)?;
         }
-    }
+        Some("unpack") => {
+            let use_bitstream = args.contains(["-b", "--bitstream"]);
 
-    let infile = infile.unwrap_or_else(|| print_help(1));
-    let outfile = outfile.unwrap_or_else(|| {
-        let mut name = infile.clone();
-        if unpack {
-            if name.extension().filter(|&e| e == "upk").is_some() {
-                name.set_extension("");
-            } else {
-                name.set_extension("bin");
-            }
-        } else {
-            let mut filename = name
-                .file_name()
-                .unwrap_or_else(|| OsStr::new(""))
-                .to_os_string();
-            filename.push(".upk");
-            name.set_file_name(filename);
+            let infile = args.free_from_os_str::<PathBuf, bool>(|s| Ok(s.into()))?;
+            let outfile = args.free_from_os_str::<PathBuf, bool>(|s| Ok(s.into()))?;
+
+            let mut data = vec![];
+            File::open(infile)?.read_to_end(&mut data)?;
+            let packed_data = upkr::unpack(&data, use_bitstream);
+            File::create(outfile)?.write_all(&packed_data)?;
         }
-        name
-    });
-
-    if config.parity_contexts != 1 && config.parity_contexts != 2 && config.parity_contexts != 4 {
-        eprintln!("--parity has to be 1, 2, or 4");
-        process::exit(1);
-    }
-
-    if !unpack {
-        let mut data = vec![];
-        File::open(infile)?.read_to_end(&mut data)?;
-        if reverse {
-            data.reverse();
+        Some(other) => {
+            bail!("Unknown subcommand '{}'", other);
         }
-
-        let mut pb = pbr::ProgressBar::new(data.len() as u64);
-        pb.set_units(pbr::Units::Bytes);
-        let mut packed_data = upkr::pack(
-            &data,
-            level,
-            config,
-            Some(&mut |pos| {
-                pb.set(pos as u64);
-            }),
-        );
-        pb.finish();
-
-        if reverse {
-            packed_data.reverse();
-        }
-
-        println!(
-            "Compressed {} bytes to {} bytes ({}%)",
-            data.len(),
-            packed_data.len(),
-            packed_data.len() as f32 * 100. / data.len() as f32
-        );
-        File::create(outfile)?.write_all(&packed_data)?;
-    } else {
-        let mut data = vec![];
-        File::open(infile)?.read_to_end(&mut data)?;
-        if reverse {
-            data.reverse();
-        }
-        let mut unpacked_data = upkr::unpack(&data, config);
-        if reverse {
-            unpacked_data.reverse();
-        }
-        File::create(outfile)?.write_all(&unpacked_data)?;
     }
 
     Ok(())
 }
 
-fn print_help(exit_code: i32) -> ! {
+fn print_help() {
     eprintln!("Usage:");
-    eprintln!("  upkr [-l level(0-9)] [config options] <infile> [<outfile>]");
-    eprintln!("  upkr -u [config options] <infile> [<outfile>]");
-    eprintln!();
-    eprintln!(" -l, --level N       compression level 0-9");
-    eprintln!(" -u, --unpack        unpack infile");
-    eprintln!();
-    eprintln!("Config options (need to match when packing/unpacking):");
-    eprintln!(" -b, --bitstream     bitstream mode");
-    eprintln!(" -p, --parity N      use N (2/4) parity contexts");
-    eprintln!(" -r, --reverse       reverse input & output");
-    process::exit(exit_code);
+    eprintln!("  upkr pack [-l level(0-9)] <infile> <outfile>");
+    eprintln!("  upkr unpack <infile> <outfile>");
+    std::process::exit(1);
 }

src/parsing_packer.rs

@@ -6,25 +6,14 @@ use crate::match_finder::MatchFinder;
 use crate::rans::{CostCounter, RansCoder};
 use crate::{lz, ProgressCallback};
 
-pub fn pack(
-    data: &[u8],
-    level: u8,
-    use_bitstream: bool,
-    parity_contexts: usize,
-    progress_cb: Option<ProgressCallback>,
-) -> Vec<u8> {
-    let mut parse = parse(
-        data,
-        Config::from_level(level),
-        parity_contexts,
-        progress_cb,
-    );
+pub fn pack(data: &[u8], level: u8, use_bitstream: bool, progress_cb: Option<ProgressCallback>) -> Vec<u8> {
+    let mut parse = parse(data, Config::from_level(level), progress_cb);
     let mut ops = vec![];
     while let Some(link) = parse {
         ops.push(link.op);
         parse = link.prev.clone();
     }
-    let mut state = lz::CoderState::new(parity_contexts);
+    let mut state = lz::CoderState::new();
     let mut coder = RansCoder::new(use_bitstream);
     for op in ops.into_iter().rev() {
         op.encode(&mut coder, &mut state);
@@ -49,7 +38,6 @@ type Arrivals = HashMap<usize, Vec<Arrival>>;
 fn parse(
     data: &[u8],
     config: Config,
-    parity_contexts: usize,
     mut progress_cb: Option<ProgressCallback>,
 ) -> Option<Rc<Parse>> {
     let mut match_finder = MatchFinder::new(data)
@@ -141,7 +129,7 @@ fn parse(
         0,
         Arrival {
             parse: None,
-            state: lz::CoderState::new(parity_contexts),
+            state: lz::CoderState::new(),
             cost: 0.0,
         },
         max_arrivals,