c unpacker now works

c_unpacker/.gitignore

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

c_unpacker/Makefile

@@ -0,0 +1,10 @@
+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/main.c

@@ -0,0 +1,25 @@
+#include <stdio.h>
+#include <stdlib.h>
+
+int 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);
+  
+  int out_size = upkr_unpack(output_buffer, input_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

@@ -0,0 +1,4 @@
+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

@@ -0,0 +1,95 @@
+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);
+        upkr_probs[context_index] = prob + ((256 - prob + 8) >> 4);
+    } else {
+        upkr_state = (256 - prob) * (upkr_state >> 8) + (upkr_state & 255) - prob;
+        upkr_probs[context_index] = prob - ((prob + 8) >> 4);
+    }
+    
+    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);
+}
+
+int 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 {
+            u8 context_index = 1;
+            for(int i = 0; i < 8; ++i) {
+                int bit = upkr_decode_bit(context_index);
+                context_index = (context_index << 1) + bit;
+            }
+            *write_ptr++ = context_index;
+            prev_was_match = 0;
+        }
+    }
+    
+    return write_ptr - (u8*)destination;
+}

src/lib.rs

@@ -21,3 +21,12 @@ pub fn pack(
         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

@@ -9,32 +9,29 @@ pub enum Op {
 
 impl Op {
     pub fn encode(&self, coder: &mut dyn EntropyCoder, state: &mut CoderState) {
-        let base_context = 256 * (state.pos & 3);
         match self {
             &Op::Literal(lit) => {
-                encode_bit(coder, state, base_context, false);
+                encode_bit(coder, state, 0, false);
                 let mut context_index = 1;
                 for i in (0..8).rev() {
                     let bit = (lit >> i) & 1 != 0;
-                    encode_bit(coder, state, base_context + context_index, bit);
+                    encode_bit(coder, state, context_index, bit);
                     context_index = (context_index << 1) | bit as usize;
                 }
-                state.pos += 1;
                 state.prev_was_match = false;
             }
             &Op::Match { offset, len } => {
-                encode_bit(coder, state, base_context, true);
+                encode_bit(coder, state, 0, true);
                 if !state.prev_was_match {
-                    encode_bit(coder, state, 1024, offset != state.last_offset);
+                    encode_bit(coder, state, 256, offset != state.last_offset);
                 } else {
                     assert!(offset != state.last_offset);
                 }
                 if offset != state.last_offset {
-                    encode_length(coder, state, 1025, offset + 1);
+                    encode_length(coder, state, 257, offset + 1);
                     state.last_offset = offset;
                 }
-                encode_length(coder, state, 1025 + 64, len);
-                state.pos += len as usize;
+                encode_length(coder, state, 257 + 64, len);
                 state.prev_was_match = true;
             }
         }
@@ -42,11 +39,11 @@ impl Op {
 }
 
 pub fn encode_eof(coder: &mut dyn EntropyCoder, state: &mut CoderState) {
-    encode_bit(coder, state, 256 * (state.pos & 3), true);
+    encode_bit(coder, state, 0, true);
     if !state.prev_was_match {
-        encode_bit(coder, state, 1024, true);
+        encode_bit(coder, state, 256, true);
     }
-    encode_length(coder, state, 1025, 1);
+    encode_length(coder, state, 257, 1);
 }
 
 fn encode_bit(
@@ -80,16 +77,14 @@ fn encode_length(
 pub struct CoderState {
     contexts: ContextState,
     last_offset: u32,
-    pos: usize,
     prev_was_match: bool,
 }
 
 impl CoderState {
     pub fn new() -> CoderState {
         CoderState {
-            contexts: ContextState::new((1 + 255) * 4 + 1 + 64 + 64),
+            contexts: ContextState::new(1 + 255 + 1 + 64 + 64),
             last_offset: 0,
-            pos: 0,
             prev_was_match: false,
         }
     }
@@ -101,7 +96,7 @@ impl CoderState {
 
 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 contexts = ContextState::new(1 + 255 + 1 + 64 + 64);
     let mut result = vec![];
     let mut offset = 0;
     let mut prev_was_match = false;
@@ -124,15 +119,14 @@ pub fn unpack(packed_data: &[u8], use_bitstream: bool) -> Vec<u8> {
     }
 
     loop {
-        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 decoder.decode_with_context(&mut contexts.context_mut(0)) {
+            if prev_was_match || decoder.decode_with_context(&mut contexts.context_mut(256)) {
+                offset = decode_length(&mut decoder, &mut contexts, 257) - 1;
                 if offset == 0 {
                     break;
                 }
             }
-            let length = decode_length(&mut decoder, &mut contexts, 1025 + 64);
+            let length = decode_length(&mut decoder, &mut contexts, 257 + 64);
             for _ in 0..length {
                 result.push(result[result.len() - offset]);
             }
@@ -141,7 +135,7 @@ pub fn unpack(packed_data: &[u8], use_bitstream: bool) -> 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(base_context + context_index));
+                let bit = decoder.decode_with_context(&mut contexts.context_mut(context_index));
                 context_index = (context_index << 1) | bit as usize;
                 byte |= (bit as u8) << i;
             }