Upload files to ""

langsmoke

@@ -0,0 +1,37 @@
+#!/bin/bash
+
+# Get input argument
+if [ $# -ne 1 ]; then
+    echo "Usage: langsmoke <file>"
+    exit 1
+fi
+
+if [ ! -r $1 ] || [ -d $1 ]; then
+    echo "Error: $1 is not a readable file"
+    exit 1
+fi
+
+# Get size of input in MB
+size=$(du -sm $1 | cut -f1)
+let "size++"
+
+src=$1
+bz3=$(bzip3 -vf $1 /dev/null 2>&1 | grep -oE "[0-9]+ bytes" | grep -oE "[0-9]+")
+
+function do_stat {
+    # 20MB cutoff.
+    l_stat=$(cat ref_$1 $src | head -c 20000000 | bzip3 -eb $size | wc -c)
+    l_ref=$2
+    sim=$(python3 -c "print(100*(1-($l_stat - min($l_ref, $bz3)) / (max($l_ref, $bz3))))")
+    echo $sim
+}
+
+echo "C: $(do_stat c 34248)"
+echo "C++: $(do_stat cpp 33606)"
+echo "Java: $(do_stat java 61196)"
+echo "6502: $(do_stat 6502 7978)"
+echo "Z80: $(do_stat z80 20182)"
+echo "APL: $(do_stat apl 29173)"
+echo "x86: $(do_stat x86 25753)"
+echo "x64: $(do_stat x64 13671)"
+echo "Shell: $(do_stat sh 44144)"

ref.md

@@ -0,0 +1,11 @@
+
+Dataset sources:
+- C: https://github.com/kspalaiologos/asm2ws, https://github.com/Mashpoe/c-hashmap, https://github.com/kspalaiologos/modern-rzip, Raylib
+- C++: https://github.com/ThePhD/sol2, LLVM libcxx
+- Java: https://github.com/kspalaiologos/cask, https://github.com/kspalaiologos/kamilalisp, https://github.com/Anuken/Mindustry, https://github.com/openjdk/jdk
+- 6502: Chess by Peter Jennings, Day Of Week by Paul Guertin, Sweet Interpreter by Apple Computer, Mini Multitasking 6502 kernel (Joachim Deboy)
+- Z80: TSE by Michael Vincent & Robin Kay, CSX:  Copyright (c) 2004, Sean McLaughlin, Megaman and Mario clones from TI Archive (TI83+, Mirage OS).
+- APL: Lambda Calculus, apl-misc-math, apl-logic, dx by Kamila Szewczyk. https://github.com/the-carlisle-group/Text2Date. https://github.com/Dyalog/Jarvis.
+- x86: Various bits of my private code (Kamila Szewczyk), Flat Assembler by Tomasz Grysztar, DOS 2.0 chkdsk by Microsoft, basicdos: https://github.com/jeffpar/basicdos
+- x64: kspalaiologos/elf-infection, https://github.com/Sakib2263/64-Bit-NASM-Assembly-Code-Examples, https://github.com/GugaDozero/asm64-tictactoe, https://github.com/nuid64/asmerver, https://github.com/CorruptedByCPU/Cyjon
+- shell: oh my zsh, autoconf, docker pi hole