TaskLang++

A Domain-Specific Language (DSL) for Task Scheduling and Automation

---

Overview

TaskLang++ is a domain-specific language (DSL) and compiler designed to simplify task scheduling and automation. Instead of writing complex cron jobs or verbose scripts, users define tasks in a clean, readable, natural-language-like syntax. The compiler reads .task files, validates them through lexical, syntactic, and semantic analysis, and outputs an execution plan.

Why TaskLang++?

Modern systems rely heavily on task scheduling — from CI/CD pipelines and cron jobs to smart assistants. General-purpose languages can express these, but they are verbose and error-prone. TaskLang++ makes scheduling declarative, readable, and safe.

---

✨ Key Technical Highlights

This project stands out by moving beyond basic parsing to implement comprehensive compiler logic:

• LALR(1) Parsing Engine: Uses Bison to enforce strict, unambiguous grammar validation.
• Deep Semantic Analysis: Goes beyond syntax to catch logical flaws before execution.
  • Graph traversal using Depth-First Search (DFS) to detect and reject circular dependencies (taskA -> taskB -> taskA).
  • Strict reference validation to guarantee AFTER clauses point to existing tasks.
  • Duplicate identifier detection to maintain a clean namespace.
• Topological Sorting Engine: The executor uses graph algorithms to mathematically sort tasks, ensuring dependencies are perfectly resolved before any execution begins.
• Robust Error Handling: Generates precise, line-accurate errors for lexical, syntax, and semantic failures.

---

DSL Scope

Supported Task Types

• Script execution – Run any shell script or Python file
• Daily recurring – Run every day at a specific time
• Weekly recurring – Run every week on a specific day and time
• One-time timed – Run once at a specific time
• Dependency-based – Run after another task completes
• Conditional – Run only if a condition (e.g., success) is met

Scheduling Mechanisms

Mechanism	Syntax Example	Description
Daily	EVERY DAY AT 06:00	Recurring daily at a fixed time
Weekly	EVERY WEEK ON MONDAY AT 08:00	Recurring weekly on a specific day
Timed	AT 15:30	One-time execution at a set time
After (dep.)	AFTER backupDB	Runs after another task finishes
Conditional	IF success	Runs only if the dependency passed

Constraints & Assumptions

• Task names must be unique across the file
• Task names must start with a letter and may contain letters, digits, or underscores
• The RUN argument must always be a string literal (in double quotes)
• Time must follow HH:MM format (24-hour clock)
• IF conditions can only follow an AFTER dependency
• Circular dependencies between tasks are not allowed

---

Token Table

Token Name	Lexeme	Description
TASK	TASK	Begins a task definition block
RUN	RUN	Specifies the command/script to execute
EVERY	EVERY	Recurring schedule keyword
DAY	DAY	Used with EVERY for daily schedules
WEEK	WEEK	Used with EVERY for weekly schedules
ON	ON	Day specifier for weekly schedules
AT	AT	Time specification keyword
AFTER	AFTER	Dependency specification
IF	IF	Conditional execution keyword
SUCCESS	success	Condition keyword (lowercase)
WEEKDAY	MONDAY...	Day of week (MON–SUN)
IDENTIFIER	[a-zA-Z_][a-zA-Z0-9_]*	Task name or variable
STRING_LITERAL	"..."	Quoted string (script/command path)
TIME	HH:MM	24-hour time literal
NUMBER	[0-9]+	Integer literal
LBRACE	{	Opens a task body
RBRACE	}	Closes a task body
SEMICOLON	;	Statement terminator

---

Formal Grammar (EBNF)

program         = { task_definition } ;

task_definition = "TASK" identifier "{" task_body "}" ;

task_body       = run_statement
                  schedule_spec
                  [ condition_spec ] ;

run_statement   = "RUN" string_literal ";" ;

schedule_spec   = daily_schedule
                | weekly_schedule
                | timed_schedule
                | after_schedule ;

daily_schedule  = "EVERY" "DAY" "AT" time ";" ;

weekly_schedule = "EVERY" "WEEK" "ON" weekday "AT" time ";" ;

timed_schedule  = "AT" time ";" ;

after_schedule  = "AFTER" identifier ";" ;

condition_spec  = "IF" condition ";" ;

condition       = "success" ;

time            = digit digit ":" digit digit ;

weekday         = "MONDAY" | "TUESDAY" | "WEDNESDAY" | "THURSDAY"
                | "FRIDAY" | "SATURDAY" | "SUNDAY" ;

identifier      = letter { letter | digit | "_" } ;

string_literal  = '"' { character } '"' ;

digit           = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9" ;

letter          = "a" | ... | "z" | "A" | ... | "Z" ;

---

Compiler Architecture

---

Example Programs

1. Simple Daily Task

TASK dailyReport {
    RUN "report.py";
    EVERY DAY AT 06:00;
}

Output:

2. Multi-Step Workflow with Dependencies

TASK backupDB {
    RUN "backup.sh";
    EVERY DAY AT 02:00;
}

TASK sendReport {
    RUN "report.py";
    AFTER backupDB;
    IF success;
}

TASK cleanup {
    RUN "cleanup.sh";
    EVERY WEEK ON SUNDAY AT 03:00;
}

Output:

3. Complex Pipeline Execution

For more advanced scenarios with multiple chained tasks, the topological sorting engine mathematically guarantees the precise order of execution across the dependency graph:

---

Project Structure

TaskLang++/
├── examples/
│   ├── error_examples/
│   │   ├── circular_deps.task     # Circular dependency example
│   │   ├── duplicate_task.task    # Duplicate task definition
│   │   ├── invalid_syntax.task    # Missing semicolons, bad syntax
│   │   ├── type_error.task        # RUN given a number, not string
│   │   └── undefined_task.task    # AFTER references non-existent task
│   ├── complex.task               # Complex multi-task scenario
│   ├── conditions.task            # Conditional execution demo
│   ├── simple_daily.task          # Basic daily schedule
│   └── workflow.task              # Multi-task dependency workflow
├── src/
│   ├── ast.c                      # AST node creation and management
│   ├── ast.h                      # AST struct definitions
│   ├── executor.c                 # Task execution engine
│   ├── executor.h                 # Executor function declarations
│   ├── lexer.l                    # Flex lexer specification
│   ├── main.c                     # Entry point – ties all phases together
│   ├── parser.y                   # Bison parser grammar
│   ├── semantic.c                 # Semantic analysis and validation
│   └── semantic.h                 # Semantic function declarations
├── .gitignore
├── build.sh                       # Linux/WSL build script
├── Makefile                       # Makefile for building the compiler
└── README.md                      # This file

---

Building the Project

Prerequisites

Tool	Purpose
gcc	C compiler
flex	Lexer generator (Lex)
bison	Parser generator (Yacc)
make	Build automation

Build (Linux / WSL)

chmod +x build.sh
./build.sh

Or with Make:

make

Run

./tasklang examples/simple_daily.task
./tasklang examples/workflow.task
./tasklang examples/error_examples/circular_deps.task

---

Semantic Validations

Check	Description
Duplicate task names	Error if two tasks share the same name
Undefined task references	Error if AFTER references a task that doesn't exist
Type checking	Error if RUN receives a number instead of a string
Circular dependencies	DFS-based cycle detection in the task dependency graph

---

🚀 Future Extensions

While fully functional within its defined scope, TaskLang++ lays the groundwork for several advanced features:

• Symbol Table Implementation: Upgrading O(N) linear scans to an O(1) Hash Map for scalable compilation of massive task files.
• Variables & Parameters: Allowing string interpolation (RUN "script.sh --user ${user}") within tasks.
• Real Execution Backend: Hooking the Topological Sorting Engine directly into cron, systemd, or schtasks to actually execute the compiled job plans on the host OS.
• Language Server Protocol (LSP): Providing IDE support (syntax highlighting, inline errors) for .task files.

---

Author

Udayaratna H.S.S