Interpreter

Pascal in the browser

Result 

                
                
              

            


            

              Runtime is lazy-loaded. First run downloads Pyodide, then later runs use the warm Python VM.
            


            

              How this runs a Python interpreter on a static website
              

                

                  The original interpreter is a Python project, but this site is hosted as static files on GitHub Pages.
                  To keep the project real without adding a backend server, this page runs Python directly in the browser
                  through Pyodide, a WebAssembly build of CPython.
                

                

                  The page first renders the editor, controls, and output panels using ordinary HTML, CSS, and JavaScript.
                  The heavier Python runtime is loaded only when the demo is likely to be used: when the demo scrolls into
                  view, or when the Run button is hovered or focused. On first load, the script downloads Pyodide, fetches
                  the interpreter modules from the GitHub repository, writes them into Pyodide's in-browser filesystem,
                  and exposes a small JavaScript-to-Python bridge.
                

                

                  That bridge sends the editor text into the Python pipeline: lexer, parser, semantic analyser, bytecode
                  emitter, and either the stack VM or the tree-walking interpreter. The result is returned to JavaScript as
                  structured JSON, then shown in the Output, AST/Bytecode, and Scope panels.
                

                

                  Demo programs can use READLN(name) for browser inputs. The page detects those calls, builds
                  the small Inputs panel, and substitutes the entered values before passing the source to the current
                  Python interpreter.
                

                

                  Lazy loading matters because Pyodide is much larger than the rest of this site. Loading it only when the
                  interpreter demo is needed keeps the homepage and normal browsing fast, while still letting the live demo
                  run the actual Python implementation rather than a rewritten imitation.