Diakrytynizator

A highly optimized x86_64 assembly program that transforms UTF-8 encoded text by applying polynomial transformations to Unicode characters.

Overview

Diakrytynizator reads UTF-8 text from standard input, applies a polynomial transformation to characters with Unicode values greater than 0x7F, and outputs the transformed text to standard output. ASCII characters (0x00-0x7F) remain unchanged.

How It Works

Polynomial Transformation

The program accepts command-line arguments that define a polynomial:

./diakrytynizator a0 a1 a2 ... an

This defines the polynomial:

w(x) = an * x^n + ... + a2 * x^2 + a1 * x + a0

Transformation rule: For each Unicode character with value x > 0x7F, the program:

1. Computes w(x - 0x80) mod 0x10FF80
2. Outputs the character with Unicode value w(x - 0x80) + 0x80

UTF-8 Validation

The program strictly validates UTF-8 encoding:

• Accepts Unicode values from 0x00 to 0x10FFFF
• Supports 1-4 byte UTF-8 sequences
• Rejects overlong encodings (only shortest form accepted)
• Returns exit code 1 on invalid input

Building

Compile with NASM and link:

nasm -f elf64 -w+all -w+error -o diakrytynizator.o diakrytynizator.asm
ld --fatal-warnings -o diakrytynizator diakrytynizator.o

Usage Examples

Identity transformation (w(x) = x):

echo "Zażółć gęślą jaźń…" | ./diakrytynizator 0 1
# Output: Zażółć gęślą jaźń…
# Exit code: 0

Constant transformation (w(x) = 133):

echo "Zażółć gęślą jaźń…" | ./diakrytynizator 133
# Output: Zaąąąą gąąlą jaąąą
# Exit code: 0

Complex polynomial (w(x) = x² + 623420x + 1075041):

echo "ŁOŚ" | ./diakrytynizator 1075041 623420 1
# Output: „O"
# Exit code: 0

Error handling (invalid UTF-8):

echo -e "abc\n\x80" | ./diakrytynizator 7
# Output: abc
# Exit code: 1

Technical Details

Implementation Features

• Buffered I/O: Uses 1KB buffers for efficient reading and writing
• Modular arithmetic: All polynomial computations performed modulo 0x10FF80
• UTF-8 parsing: Hand-optimized byte-level UTF-8 decoding and encoding
• Parameter validation: Validates polynomial coefficients (non-negative integers, no leading zeros)
• Error handling: Comprehensive validation of input encoding and parameters

Architecture

The program consists of several key components:

• parse: Converts decimal string arguments to integers with validation
• convert_args: Processes command-line arguments into polynomial coefficients
• calc_poly: Evaluates the polynomial using Horner's method
• utf_count_bytes: Determines UTF-8 character byte length from first byte
• utf_bytes_for_code: Calculates required bytes for a Unicode value
• load_head_byte/load_tail_byte: UTF-8 decoder
• add_char: UTF-8 encoder that writes transformed characters to output buffer

Memory Layout

• .data: Constants for UTF-8 masks, prefixes, and modulo value
• .bss: Input/output buffers (1KB each)
• Stack-based polynomial coefficient storage

Exit Codes

• 0: Successful execution
• 1: Error (invalid parameters, malformed UTF-8, or encoding violation)

Acknowledgments

This program was developed as an assignment for the Operating Systems course at the University of Warsaw in 2021.