file names adjusted

.github/workflows/build_wheels.yml

@@ -22,7 +22,7 @@ jobs:
             gcc-12-riscv64-linux-gnu
 
       - name: Build object files
-        run: bash src/copapy/obj/crosscompile.sh
+        run: bash tools/crosscompile.sh
 
       - uses: actions/upload-artifact@v4
         with:

.github/workflows/ci.yml

@@ -13,7 +13,7 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Build object files
-        run: bash src/copapy/obj/nativecompile.sh
+        run: bash tools/nativecompile.sh
 
       - uses: actions/upload-artifact@v4
         with:

.gitignore

@@ -10,7 +10,6 @@ __pycache__
 .pytest_cache
 /tests/autogenerated_*.py
 *test*.o
-/bin/*
 token.txt
 /src/copapy/obj/*.o
 /src/copapy/obj/old/*.o
@@ -20,5 +19,4 @@ bin/*
 build/*
 /*.obj
 /src/*.pyd
-src/copapy/stencils.c
 vc140.pdb

inspect.sh

@@ -1,6 +0,0 @@
-#!/bin/bash
-
-source build.sh
-python tests/test_compile_div.py
-python tools/extract_code.py
-objdump -D -b binary -m i386:x86-64 --adjust-vma=0x1000 bin/test_code.bin > bin/test_code.bin.txt

src/copapy/stencils.c

@@ -0,0 +1,291 @@
+
+    // Auto-generated stencils for copapy
+    // Do not edit manually
+
+    volatile int dummy_int = 1337;
+    volatile float dummy_float = 1337;
+    
+    void result_int(int arg1);
+    
+    void result_float(float arg1);
+    
+    void result_int_int(int arg1, int arg2);
+    
+    void result_int_float(int arg1, float arg2);
+    
+    void result_float_int(float arg1, int arg2);
+    
+    void result_float_float(float arg1, float arg2);
+    
+    void add_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 + arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void add_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 + arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void add_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 + arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void add_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 + arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void sub_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 - arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void sub_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 - arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void sub_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 - arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void sub_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 - arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void mul_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 * arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void mul_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 * arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void mul_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 * arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void mul_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 * arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void div_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int((float)arg1 / arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void div_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float((float)arg1 / arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void div_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 / arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void div_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 / arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void gt_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 > arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void gt_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 > arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void gt_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 > arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void gt_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 > arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void eq_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 == arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void eq_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 == arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void eq_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1 == arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void eq_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1 == arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void mod_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1 % arg2, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg0_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(dummy_int, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg1_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1, dummy_int);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg0_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(dummy_float, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg1_int_int(int arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_float(arg1, dummy_float);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg0_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_float(dummy_int, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg1_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(arg1, dummy_int);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg0_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(dummy_float, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg1_int_float(int arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_float(arg1, dummy_float);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg0_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_int(dummy_int, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg1_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1, dummy_int);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg0_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(dummy_float, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg1_float_int(float arg1, int arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1, dummy_float);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg0_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_int_float(dummy_int, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_int_reg1_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_int(arg1, dummy_int);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg0_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(dummy_float, arg2);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void read_float_reg1_float_float(float arg1, float arg2) {
+        asm volatile (".long 0xE1401F0F");
+        result_float_float(arg1, dummy_float);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void write_int(int arg1) {
+        asm volatile (".long 0xE1401F0F");
+        dummy_int = arg1;
+        result_int(arg1);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    void write_float(float arg1) {
+        asm volatile (".long 0xE1401F0F");
+        dummy_float = arg1;
+        result_float(arg1);
+        asm volatile (".long 0xE2401F0F");
+    }
+    
+    int function_start(){
+        result_int(0);  // dummy call instruction before marker gets striped
+        asm volatile (".long 0xE2401F0F");
+        return 1;
+    }
+    
+    int function_end(){
+        result_int(0);
+        asm volatile (".long 0xE1401F0F");
+        return 1;
+    }
+    

test.sh

@@ -1,7 +0,0 @@
-#!/bin/bash
-set -e
-echo "Compile..."
-python tests/test_compile.py
-echo "Run..."
-echo "-----------------------------------"
-bin/coparun bin/test.copapy

tools/build.sh

@@ -1,9 +1,9 @@
 #!/bin/bash
 set -e
 set -v
-SRC=src/copapy/stencils.c
+SRC=bin/stencils.c
 DEST=src/copapy/obj
-python src/copapy/generate_stencils.py $SRC
+python tools/generate_stencils.py $SRC
 mkdir -p $DEST
 gcc -c $SRC -O0 -o $DEST/stencils_x86_64_O0.o
 gcc -c $SRC -O1 -o $DEST/stencils_x86_64_O1.o

tools/crosscompile.sh

@@ -13,18 +13,18 @@
 set -e
 set -v
 
-SRC=src/copapy/stencils.c
+SRC=bin/stencils.c
 DEST=src/copapy/obj
 OPT=O3
 
 mkdir -p $DEST
 
 # Windows x86_64 (ARM64)
-python src/copapy/generate_stencils.py --abi ms $SRC
+python tools/generate_stencils.py --abi ms $SRC
 gcc-12 -$OPT -c $SRC -o $DEST/stencils_AMD64_$OPT.o
 
 # Native x86_64
-python src/copapy/generate_stencils.py $SRC
+python tools/generate_stencils.py $SRC
 gcc-12 -$OPT -c $SRC -o $DEST/stencils_x86_64_$OPT.o
 
 # ARM64

tools/extract_code.py

@@ -1,22 +1,28 @@
 from copapy.binwrite import data_reader, Command, ByteOrder
 from copapy.stencil_db import RelocationType
+import argparse
 
-input_file = "bin/test.copapy"
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("input_file", type=str, help="Input file path with copapy commands")
+    parser.add_argument("output_file", type=str, help="Output file with patched code")
+    parser.add_argument("--data_section_offset", type=int, default=0x2000, help="Offset for data relative to code section")
+    parser.add_argument("--byteorder", type=str, choices=['little', 'big'], default='little', help="Select byteorder")
+    args = parser.parse_args()
 
-output_file = "bin/test_code.bin"
+    input_file: str = args.input_file
+    output_file: str = args.output_file
+    data_section_offset: int = args.data_section_offset
+    byteorder: ByteOrder = args.byteorder
 
-data_section_offset = 0x2000
+    with open(input_file, mode='rb') as f:
-
-byteorder: ByteOrder = 'little'
-
-with open(input_file, mode='rb') as f:
         dr = data_reader(f.read(), byteorder)
 
-buffer_index: int = 0
+    buffer_index: int = 0
-end_flag: int = 0
+    end_flag: int = 0
-program_data: bytearray = bytearray([])
+    program_data: bytearray = bytearray([])
 
-while (end_flag == 0):
+    while (end_flag == 0):
         com = dr.read_com()
 
         if com == Command.ALLOCATE_DATA:
@@ -64,7 +70,7 @@ while (end_flag == 0):
         else:
             assert False, f"Unknown command: {com}"
 
-with open(output_file, mode='wb') as f:
+    with open(output_file, mode='wb') as f:
         f.write(program_data)
 
-print('OK')
+    print(f"Code written to {output_file}.")

tools/inspect.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+
+source tools/build.sh
+python tests/test_compile_div.py
+python tools/extract_code.py "bin/test.copapy" "bin/test.copapy.bin"
+objdump -D -b binary -m i386:x86-64 --adjust-vma=0x1000 bin/test.copapy.bin > bin/test.copapy.asm

tools/nativecompile.sh

@@ -2,10 +2,10 @@
 set -e
 set -v
 
-SRC=src/copapy/stencils.c
+SRC=bin/stencils.c
 DEST=src/copapy/obj
 
-python src/copapy/generate_stencils.py $SRC
+python tools/generate_stencils.py $SRC
 
 mkdir -p $DEST
 gcc-12 -c $SRC -O0 -o $DEST/stencils_x86_64_O0.o
@@ -13,7 +13,7 @@ gcc-12 -c $SRC -O1 -o $DEST/stencils_x86_64_O1.o
 gcc-12 -c $SRC -O2 -o $DEST/stencils_x86_64_O2.o
 gcc-12 -c $SRC -O3 -o $DEST/stencils_x86_64_O3.o
 
-python src/copapy/generate_stencils.py --abi ms $SRC
+python tools/generate_stencils.py --abi ms $SRC
 
 gcc-12 -c $SRC -O0 -o $DEST/stencils_AMD64_O0.o
 gcc-12 -c $SRC -O1 -o $DEST/stencils_AMD64_O1.o