NodeScript 1.2.2

NodeScript

NodeScript is a rudimentary programming language designed to function on 'nodes', intended for use in puzzle games. Find more details here

Nodes

• Input nodes will continuously attempt to send the next line. It has only one output.
• Regular nodes can take in one string at a time, process it and send it to a set number of outputs. Each node can store a single string in mem. mem is the only variable which will persist between executions.
• Combiner nodes can merge multiple inputs into one output. It offers no options to choose which string goes through first, picking whatever comes first.
• Output nodes will consume the lines sent to it, storing it in a string.

Features

• Basic arithmetic and boolean logic
• Variables - All variables are global in scope
• Indexing
  • Element_of: v[i]
  • Slice: v[i:j]
• Basic control flow (if-else)
• Native functions for things like:
  • String manipulation
  • Data conversion and parsing
• Dynamic typing between:
  • string
  • string[]
  • int
  • bool

Notably, there are NO loops within the scripting itself. No while. No for. Execution will occur line by line and will only start when a node receives an input string to be processed.

Syntax

Every line contains a single statement. All statements will start with a relevant keyword for the operation.

• SET: Sets a variable to a certain value. Variables do not need to be declared. Syntax: SET <variable_name>, <expression>
• PRINT: Sends a string to a specific output node, denoted by an index. Syntax: PRINT <output_idx>, <expression>
• RETURN: Ends the program (until the next input comes). Syntax: RETURN
• IF: Executes the following code if the given expression is true. Syntax IF <expression>
• ELSE: Executes the following code if the previous if statement was false. Syntax ELSE
• ENDIF: Marks the end of the IF clause. Either ends the IF code section or the ELSE code section. Only one is needed per IF/ELSE statement. Syntax ENDIF
• NOP: Does nothing. Helpful for synchronizing the timing of multiple nodes. Syntax NOP

Benchmarks

NodeScriptBenchmark measures each phase of the compile pipeline (tokenize/parse, optimize/validate/compile), cold vs. warm execution, Reset, and complete graph runs across several representative workloads (straight-line arithmetic, branch-heavy code, native calls, string/array operations, compile errors, and runtime errors, each in tiny/medium/max sizes) and graph shapes (linear, fan-out, fan-in, disconnected, and cycle). All benchmark inputs are deterministic and checked into NodeScriptBenchmark/TestData. Results report mean/median, throughput, allocated bytes and GC collections via MemoryDiagnoser, alongside the host runtime, architecture and GC settings that BenchmarkDotNet records automatically. A baseline run is recorded in NodeScriptBenchmark/Baseline/README.md.

• CI-safe subset (fast, used by the Benchmark workflow): dotnet run -c Release --project ./NodeScriptBenchmark
• Full/local run (more iterations, statistically robust): dotnet run -c Release --project ./NodeScriptBenchmark -- --job Medium
• A subset of benchmarks: dotnet run -c Release --project ./NodeScriptBenchmark -- --filter *PipelineBenchmarks*

Development and CI details are documented in the wiki.