add --version flag to output upkr version

Cargo.lock

@@ -172,7 +172,7 @@ checksum = "dcc811dc4066ac62f84f11307873c4850cb653bfa9b1719cee2bd2204a4bc5dd"
 
 [[package]]
 name = "upkr"
-version = "0.2.0-pre3"
+version = "0.2.0"
 dependencies = [
  "anyhow",
  "cdivsufsort",

Cargo.toml

@@ -1,9 +1,10 @@
 [package]
 name = "upkr"
-version = "0.2.0-pre3"
+version = "0.2.0"
 edition = "2021"
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+[profile.release]
+strip = "debuginfo"
 
 [dependencies]
 cdivsufsort = "2"

README.md

@@ -2,13 +2,57 @@
 
 Upkr is a simple general purpose lz packer designed to be used in the [MicroW8](https://github.com/exoticorn/microw8) platform.
 The compressed format is losely based on [Shrinkler](https://github.com/askeksa/Shrinkler) with the main difference being that
-Upkr doesn't differnetiate between literals at odd or even addresses and that I went with rANS/rABS instead of a range coder.
+Upkr doesn't differentiate between literals at odd or even addresses (by default) and that I went with rANS/rABS instead of a range coder.
 
-At this point, Upkr should still be considered unstable - the compressed format is not very likely to change but I still want
+Compression rate is on par with Shrinkler.
-to keep that option open a little longer.
+
+The differences compare to Shrinkler also makes it interesting on 8bit platforms. The z80 unpacker included in the release
+is both about twice as fast and smaller than the Shrinkler unpacker.
 
 ## Inspirations:
 
 * Ferris' blog about his [C64 intro packer](https://yupferris.github.io/blog/2020/08/31/c64-4k-intro-packer-deep-dive.html)
 * [Shrinkler](https://github.com/askeksa/Shrinkler)
 * Ryg's [sample rANS implementation](https://github.com/rygorous/ryg_rans)
+
+## Unpackers
+
+The release includes a reference c unpacker, as well as some optimized asm unpackers (arm and riscv). The unpckers in
+c_unpacker and asm_unpackers unpack the default upkr compressed format. The z80_unpacker
+is based on some variations to the compressed format. (Use `upkr --z80` to select those variations.)
+An optimized x86 (DOS) unpacker is currently being worked on out of tree.
+
+## Usage
+
+```
+  upkr [-l level(0-9)] [config options] <infile> [<outfile>]
+  upkr -u [config options] <infile> [<outfile>]
+  upkr --margin [config options] <infile>
+
+ -l, --level N       compression level 0-9
+ -0, ..., -9         short form for setting compression level
+ -u, --unpack        unpack infile
+ --margin            calculate margin for overlapped unpacking of a packed file
+
+Config presets for specific unpackers:
+ --z80               --big-endian-bitstream --invert-bit-encoding --simplified-prob-update -9
+ --x86               --bitstream --invert-is-match-bit --invert-continue-value-bit --invert-new-offset-bit
+ --x86b              --bitstream --invert-continue-value-bit --no-repeated-offsets -9
+
+Config options (need to match when packing/unpacking):
+ -b, --bitstream     bitstream mode
+ -p, --parity N      use N (2/4) parity contexts
+ -r, --reverse       reverse input & output
+
+Config options to tailor output to specific optimized unpackers:
+ --invert-is-match-bit
+ --invert-new-offset-bit
+ --invert-continue-value-bit
+ --invert-bit-encoding
+ --simplified-prob-update
+ --big-endian-bitstream   (implies --bitstream)
+ --no-repeated-offsets
+ --eof-in-length
+ --max-offset N
+ --max-length N
+```

asm_unpackers/Makefile

@@ -39,6 +39,19 @@ build/unpack_armv6m.bin: unpack_armv6m.S
 	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_arm32: ../c_unpacker/main.c unpack_arm32.S
+	mkdir -p build
+	arm-linux-gnueabihf-gcc -g -static -o $@ $^
+
+test_arm32: build/unpack_arm32
+	qemu-arm $< test_data.upk /tmp/out.bin
+	cmp test_data.bin /tmp/out.bin
+
+build/unpack_arm32.bin: unpack_arm32.S
+	mkdir -p build
+	arm-none-eabi-gcc -c -o build/unpack_arm32.o $?
+	arm-none-eabi-objcopy -O binary --only-section=.text build/unpack_arm32.o $@
+	
 build/unpack_c: ../c_unpacker/main.c ../c_unpacker/unpack.c
 	mkdir -p build
 	gcc -g -o $@ $^
@@ -47,5 +60,5 @@ 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 build/unpack_riscv32.bin
+sizes: build/unpack_armv6m.bin build/unpack_riscv64.bin build/unpack_riscv32.bin build/unpack_arm32.bin
 	ls -l build/*.bin

asm_unpackers/unpack_arm32.S

@@ -0,0 +1,100 @@
+.arm
+
+.section .text
+
+.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 .. context index
+// r6 .. decode_length temp
+// r7 .. probs ptr
+// r8-r11 .. decode_bit temp
+// r12 .. decode_length return address
+upkr_unpack:
+    push { r3-r11, lr }
+
+	mov r2, #384
+    mov r3, #128
+.Lclear:
+    subs r2, r2, #1
+    strb r3, [sp, -r2]
+    bne .Lclear
+
+.Lloop:
+    mov r5, #0
+    bl upkr_decode_bit
+    bcc .Ldata
+.Lmatch:
+	mov r5, #256
+    rsbs r6, r4, #0
+    blcc upkr_decode_bit
+    bcc .Lskip_offset
+
+    bl upkr_decode_length
+    adds r3, r4, #1
+    popeq { r3-r11, pc }
+.Lskip_offset:
+
+	mov r5, #256+64
+    bl upkr_decode_length
+.Lcopy_loop:
+    ldrb r5, [r0, r3]
+.Lstore:
+    strb r5, [r0], #1
+    adds r4, r4, #1
+    blt .Lcopy_loop
+    b .Lloop
+
+.Ldata:
+    mov r5, #1
+
+.Ldata_loop:
+    bl upkr_decode_bit
+    adc r5, r5, r5
+	movs r4, r5, lsr #8
+    beq .Ldata_loop
+	b .Lstore
+
+.type upkr_decode_length, %function
+upkr_decode_length:
+	mov r12, lr
+
+    mov r4, #0
+	mvn r6, #0
+.Lbit_loop:
+    bl upkr_decode_bit_inc
+	addcc r4, r4, r6
+	movcc pc, r12
+
+    bl upkr_decode_bit_inc
+    addcs r4, r4, r6
+	mov r6, r6, lsl #1
+    b .Lbit_loop
+
+.type upkr_decode_bit, %function
+upkr_decode_bit_inc:
+	add r5, r5, #1
+upkr_decode_bit:
+	cmp r2, #4096
+	ldrltb r8, [r1], #1
+	orrlt r2, r8, r2, lsl#8
+	blt upkr_decode_bit
+
+    ldrb r8, [sp, -r5]
+	and r9, r2, #255
+	add r9, r9, #1
+	cmp r8, r9
+	rsbcs r8, r8, #256
+	mvn r9, r2, lsr#8
+	addcs r9, r9, #1
+	mla r2, r8, r9, r2
+	add r9, r8, #8
+	sub r8, r8, r9, lsr#4
+	rsbcs r8, r8, #256
+	strb r8, [sp, -r5]
+	mov pc, r14
+

asm_unpackers/unpack_riscv.S

@@ -1,13 +1,11 @@
 .section .text
 
-#define FRAME_SIZE (256+32*4+4)
-
-// x8 prob array ptr
 // x9 prev was literal
 // x10 out ptr
 // x11 in ptr
 // x12 offset
 // x13 state
+// x14 context index
 
 .global upkr_unpack
 .type upkr_unpack, %function
@@ -15,11 +13,11 @@ upkr_unpack:
 	mv t4, ra
 	mv x17, x8
 	mv t6, x9
-	li x13, FRAME_SIZE
+	li x9, 256 + 128
-	li x9, 128
+	mv x13, x9
 1:
-	addi sp, sp, -1
+	sub x8, sp, x13
-	sb x9, 0(sp)
+	sb x9, 0(x8)
 	addi x13, x13, -1
 	bnez x13, 1b
 	
@@ -35,7 +33,7 @@ upkr_unpack:
 
 .Lfinished_offset:
 	addi x14, x14, 64
-	jal t3, upkr_decode_number
+	jalr ra // jal upkr_decode_number
 1:
 	add x14, x10, t0
 	lbu x14, (x14)
@@ -58,36 +56,14 @@ upkr_unpack:
 .Lread_offset_inc_x14:
 	addi x14, x14, 1
 .Lread_offset:
-	jal t3, upkr_decode_number
+	jalr ra // jal upkr_decode_number
 	addi t0, x9, 1
 	bnez t0, .Lfinished_offset
 .Ldone:
-	addi sp, sp, FRAME_SIZE
 	mv x8, x17
 	mv x9, t6
 	jr t4
 
-// x14 context index
-// return: x9 negtive decoded number 
-upkr_decode_number:
-	mv t5, x14
-	li x9, 0
-	li x8, -1
-1:
-	jal upkr_decode_bit
-	beqz x15, 1f
-	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
@@ -104,39 +80,52 @@ upkr_decode_bit:
 	srli x15, x13, 12
 	beqz x15, upkr_load_byte
 
-	mv t1, x14
+	addi x14, x14, 1
-	mv t2, x10
 	
-	add x14, x14, sp
+	sub t2, sp, x14
-	lbu x12, 0(x14)
+	lbu x12, (t2)
 	
-	andi x10, x13, 255
+	andi x8, x13, 255
-	sltu x15, x10, x12
+	sltu x15, x8, x12
-	srli x13, x13, 8
+	beqz x15, 1f
-	beqz x15, .Lelse
+	xori x12, x12, 255
+	addi x12, x12, 1
+1:
+	srli x8, x13, 8
+	addi x8, x8, 1
+	sub x8, x8, x15
+	mul x8, x8, x12
+	sub x13, x13, x8
 	
-	mul x13, x13, x12
+	addi x8, x12, 8
-	add x13, x13, x10
+	srli x8, x8, 4
-	li x10, 256 + 8
+	sub x12, x12, x8
-	sub x10, x10, x12
+	beqz x15, 1f
-	srli x10, x10, 4
+	sub x12, x0, x12
-	add x12, x12, x10
+1:
-	j .Lendif
 
-.Lelse:
+	sb x12, (t2)
-	li x16, 256
-	sub x16, x16, x12
-	mul x13, x13, x16
-	add x13, x13, x10
-	sub x13, x13, x12
-	addi x10, x12, 8
-	srli x10, x10, 4
-	sub x12, x12, x10
 	
-.Lendif:
+	jalr ra
 
-	sb x12, 0(x14)
+// x14 context index
+// return: x9 negtive decoded number 
+upkr_decode_number:
+	mv t3, ra
+	mv t5, x14
+	li x9, 0
+	li t1, -1
+1:
+	jal upkr_decode_bit
+	beqz x15, 1f
+	jal upkr_decode_bit
+	beqz x15, 2f
+	add x9, x9, t1
+2:
+	add t1, t1, t1
+	j 1b
+1:
+	add x9, x9, t1
 	
-	addi x14, t1, 1
+	mv x14, t5
-	mv x10, t2
+	jr t3
-	ret

release/.gitignore

@@ -0,0 +1,4 @@
+*.zip
+*.tgz
+upkr-linux/
+upkr-windows/

release/Makefile

@@ -0,0 +1,35 @@
+VERSION := $(shell cargo run --release -- --version)
+
+all: clean upkr-linux-$(VERSION).tgz upkr-windows-$(VERSION).zip
+
+clean:
+	rm -rf upkr-linux
+	rm -f upkr-linux*.tgz
+	rm -rf upkr-windows
+	rm -f upkr-windows*.zip
+
+upkr-linux-$(VERSION).tgz: upkr-linux/upkr PHONY
+	cp ../README.md upkr-linux
+	cd .. && git archive HEAD c_unpacker | tar -xC release/upkr-linux
+	cd .. && git archive HEAD z80_unpacker | tar -xC release/upkr-linux
+	cd .. && git archive HEAD asm_unpackers | tar -xC release/upkr-linux
+	tar czf $@ upkr-linux
+
+upkr-windows-$(VERSION).zip: upkr-windows/upkr.exe PHONY
+	cp ../README.md upkr-windows/
+	cd .. && git archive HEAD c_unpacker | tar -xC release/upkr-windows
+	cd .. && git archive HEAD z80_unpacker | tar -xC release/upkr-windows
+	cd .. && git archive HEAD asm_unpackers | tar -xC release/upkr-windows
+	zip -r -9 $@ upkr-windows
+
+upkr-linux/upkr:
+	cargo build --target x86_64-unknown-linux-musl --release
+	mkdir -p upkr-linux
+	cp ../target/x86_64-unknown-linux-musl/release/upkr upkr-linux/
+
+upkr-windows/upkr.exe:
+	cargo build --target x86_64-pc-windows-gnu --release
+	mkdir -p upkr-windows
+	cp ../target/x86_64-pc-windows-gnu/release/upkr.exe upkr-windows/
+	
+PHONY:

src/main.rs

@@ -49,12 +49,22 @@ fn main() -> Result<()> {
                 config.is_match_bit = false;
                 config.new_offset_bit = false;
             }
+            Long("x86b") => {
+                config.use_bitstream = true;
+                config.continue_value_bit = false;
+                config.no_repeated_offsets = true;
+                level = 9;
+            }
 
             Short('u') | Long("unpack") => unpack = true,
             Long("margin") => calculate_margin = true,
             Short('l') | Long("level") => level = parser.value()?.parse()?,
             Short(n) if n.is_ascii_digit() => level = n as u8 - b'0',
             Short('h') | Long("help") => print_help(0),
+            Long("version") => {
+                println!("{}", env!("CARGO_PKG_VERSION"));
+                process::exit(0);
+            }
             Long("max-unpacked-size") => max_unpacked_size = parser.value()?.parse()?,
             Value(val) if infile.is_none() => infile = Some(val.try_into()?),
             Value(val) if outfile.is_none() => outfile = Some(val.try_into()?),
@@ -149,11 +159,16 @@ fn print_help(exit_code: i32) -> ! {
     eprintln!(" -u, --unpack        unpack infile");
     eprintln!(" --margin            calculate margin for overlapped unpacking of a packed file");
     eprintln!();
+    eprintln!("Version: {}", env!("CARGO_PKG_VERSION"));
+    eprintln!();
     eprintln!("Config presets for specific unpackers:");
     eprintln!(" --z80               --big-endian-bitstream --invert-bit-encoding --simplified-prob-update -9");
     eprintln!(
         " --x86               --bitstream --invert-is-match-bit --invert-continue-value-bit --invert-new-offset-bit"
     );
+    eprintln!(
+        " --x86b              --bitstream --invert-continue-value-bit --no-repeated-offsets -9"
+    );
     eprintln!();
     eprintln!("Config options (need to match when packing/unpacking):");
     eprintln!(" -b, --bitstream     bitstream mode");