BlitzLLVM/tools/copyright.js

// AUTOGENERATED COPYRIGHT HEADER START
// Copyright (C) 2024 Michael Fabian 'Xaymar' Dirks <info@xaymar.com>
// AUTOGENERATED COPYRIGHT HEADER END
const ignoreList = [
	/^\.git$/gi,
	/^cmake\/clang$/gi,
	/^cmake\/version$/gi,
	/^third-party$/gi,
]
const sectionStart = "AUTOGENERATED COPYRIGHT HEADER START";
const sectionEnd = "AUTOGENERATED COPYRIGHT HEADER END";
const formatStyleList = {
	"#": {
		files: [
			"cmakelists.txt"
		], exts: [
			".clang-tidy",
			".clang-format",
			".cmake",
			".editorconfig",
			".gitignore",
			".gitmodules",
			".yml",
			".sh"
		],
		prepend: [
			`# ${sectionStart}`,
		],
		append: [
			`# ${sectionEnd}`,
		],
		prefix: "# ",
		suffix: "",
	},
	";": {
		files: [
		], exts: [
			".iss",
			".iss.in",
			".bb",
			".b3d",
			".b2d",
			".bpl"
		],
		prepend: [
			`; ${sectionStart}`,
		],
		append: [
			`; ${sectionEnd}`,
		],
		prefix: "; ",
		suffix: "",
	},
	"//": {
		files: [
		], exts: [
			".c",
			".c.in",
			".cpp",
			".cpp.in",
			".h",
			".h.in",
			".hpp",
			".hpp.in",
			".js",
			".rc",
			".rc.in",
			".effect"
		],
		prepend: [
			`// ${sectionStart}`,
		],
		append: [
			`// ${sectionEnd}`,
		],
		prefix: "// ",
		suffix: "",
	},
	"<!---->": {
		files: [
		], exts: [
			".htm",
			".htm.in",
			".html",
			".html.in",
			".xml",
			".xml.in",
			".plist",
			".plist.in",
			".pkgproj",
			".pkgproj.in",
			".md"
		],
		prepend: [
			`<!-- ${sectionStart} -->`,
		],
		append: [
			`<!-- ${sectionEnd} -->`,
		],
		prefix: "<!-- ",
		suffix: " -->",
	}
};
const commitFormat="%aI|%aN <%aE>"
const debug = false;

// --------------------------------------------------------------------------------
// End of easily modified area, best not touch things beyond here.
// --------------------------------------------------------------------------------
const CHILD_PROCESS = require("node:child_process");
const PROCESS = require("node:process");
const PATH = require("node:path");
const FS = require("node:fs");
const FSPROMISES = require("node:fs/promises");
const OS = require("os");
const READLINE = require('node:readline');

if (!debug)
	console.debug = function() {}

class RateLimiter {
    constructor(limit) {
        Object.defineProperty(this, "_maximum", {
            enumerable: true,
            configurable: true,
            writable: true,
            value: 0
        });
        Object.defineProperty(this, "_available", {
            enumerable: true,
            configurable: true,
            writable: true,
            value: 0
        });
        Object.defineProperty(this, "_instances", {
            enumerable: true,
            configurable: true,
            writable: true,
            value: void 0
        });
        if (!limit) {
            this._maximum = Math.ceil(Math.max(1, OS.cpus().length / 3 * 2));
        }
        else {
            this._maximum = limit;
        }
        this._available = this._maximum;
        this._instances = [];
    }
    async queue(executor, ...args) {
        while (this._available == 0) {
            await Promise.race(this._instances);
        }
        --this._available;
        const data = {};
        data.task = new Promise((resolve, reject) => {
            try {
				if (executor.constructor.name == "Promise") {
					executor.then((res) => {
						resolve(res);
					}, (err) => {
						reject(err);
					});
                } else if (executor.constructor.name == "AsyncFunction") {
                    executor(...args).then((res) => {
                        resolve(res);
                    }, (err) => {
                        reject(err);
                    });
                }
                else {
                    resolve(executor(...args));
                }
            }
            catch (ex) {
				console.error(executor, executor.toString())
                reject(ex);
            }
        });
        data.solver = data.task.finally(() => {
            const taskIndex = this._instances.indexOf(data.task);
            if (taskIndex >= 0) {
                this._instances.splice(taskIndex, 1);
            }
            const solverIndex = this._instances.indexOf(data.solver);
            if (solverIndex >= 0) {
                this._instances.splice(solverIndex, 1);
            }
            ++this._available;
        });
        this._instances.push(data.solver);
        return await data.solver;
    }
}

let abortAllWork = false;
let gitRL = new RateLimiter(3);
let workRL = new RateLimiter();
let gitCurrentFiles;
let gitUserName;
let gitUserMail;
let gitSubmodules;
let gitDate = (new Date()).toISOString();

/** Run a process asynchronously, returning an array of messages.
 * 
 * @param {*} path 
 * @param {*} args 
 * @param {*} options 
 * @returns [ErrorCode, [{type, data, text}, ...]]
 */
async function runProcessAsync(path, args, options) {
	console.debug(arguments.callee.name, Array.from(arguments));
	return await new Promise((resolve, reject) => {
		let messages = [];

		try {
			let proc = CHILD_PROCESS.spawn(path, args, Object.assign({
				"cwd": PROCESS.cwd(),
				"encoding": "utf8",
			}, options));

			let pushOutput = function(type, chunk) {
				if (!chunk) return;

				let lastMessage = undefined;
				if (messages.length == 0 || messages[messages.length - 1].type !== type) {
					lastMessage = {
						type: type,
						chunks: [],
						length: 0,
					};
					messages.push(lastMessage);
				} else {
					lastMessage = messages[messages.length - 1];
				}
				lastMessage.length += chunk.length;
				lastMessage.chunks.push(chunk);
			};

			proc.stderr.on('data', (chunk) => { pushOutput("error", chunk); });
			proc.stderr.on('close', () => { pushOutput("output", proc.stderr.read()); });
			proc.stdout.on('data', (chunk) => { pushOutput("output", chunk); });
			proc.stdout.on('close', () => { pushOutput("output", proc.stdout.read()); });
			proc.on('error', (err) => {
				reject(err);
			})
			proc.on('exit', (code) => {
				// Merge chunks of messages into a single message.
				for (let message of messages) {
					message.data = Buffer.alloc(message.length);
					let offset = 0;
					for (let chunk of message.chunks) {
						Buffer.from(chunk).copy(message.data, offset, 0);
						offset += chunk.byteLength;
					}
					message.chunks = undefined;
					message.length = undefined;
					message.text = message.data.toString();
				}
				resolve([code, messages]);
			});
		} catch (ex) {
			reject(ex);
		}
	});
}

async function git_isIgnored(path) {
	console.debug(arguments.callee.name, Array.from(arguments));
	let rpath = PATH.relative(process.cwd(), path).replaceAll(PATH.sep, PATH.posix.sep);

	// Check manual ignore list.
	for (let ignore of ignoreList) {
		if (ignore instanceof RegExp) {
			if (ignore.global) {
				let matches = rpath.matchAll(ignore);
				for (let match of matches) {
					return true;
				}
			} else {
				if (rpath.match(ignore) !== null) {
					return true;
				}
			}
		} else if (rpath.startsWith(ignore)) {
			return true;
		}
	}

	// Check if this happens to be (in) a submodule.
	let modules = await git_subModules()
	for (let module of modules) {
		if (rpath.startsWith(module)) {
			return true;
		}
	}

	// Finally check if git ignores this file.
	let result = await gitRL.queue(runProcessAsync, "git", ["check-ignore", path], {});
	return (result[0] == 0)
}

async function git_getCurrentAuthor() {
	console.debug(arguments.callee.name, Array.from(arguments));
	if (!gitUserName) {
		let result = await gitRL.queue(runProcessAsync, "git", ["config", "user.name"], {});
		gitUserName = result[1][0].text.split('\n')[0];
	}
	if (!gitUserMail) {
		let result = await gitRL.queue(runProcessAsync, "git", ["config", "user.email"], {});
		gitUserMail = result[1][0].text.split('\n')[0];
	}
	return commitFormat.replace("%aI", gitDate).replace("%aN", gitUserName).replace("%aE", gitUserMail);
}

async function git_subModules() {
	if (!gitSubmodules) {
		gitSubmodules = new Set();

		let gitmodules = PATH.join(process.cwd(), ".gitmodules");
		if ((await FSPROMISES.stat(gitmodules)).isFile()) {
			let gmfs = FS.createReadStream(gitmodules);
			const rl = READLINE.createInterface({
				input: gmfs,
				crlfDelay: Infinity
			});
			for await(const line of rl) {
				let parts = line.trim().split("=");
				if (parts.length > 1) {
					if(parts[0].trim() == "path") {
						gitSubmodules.add(parts[1].trim());
					}
				}
			}
		}
	}

	return gitSubmodules;
}

async function git_isInCurrentCommit(file) {
	console.debug(arguments.callee.name, Array.from(arguments));
	if (!gitCurrentFiles) {
		let files = [];
		let result = await gitRL.queue(runProcessAsync, "git", ["diff", "--name-only", "--cached"], {});
		if (result[0] != 0) {
			let output = "";
			for (let message of result[1]) {
				output += message.text;
			}
			throw new Error(output);
		} else {
			lines = "";
			for (let message of result[1]) {
				if (message.type == "output")
					lines += message.text;
			}
			files = lines.split(lines.indexOf("\r\n") >= 0 ? "\r\n" : "\n");
		}

		gitCurrentFiles = new Set(files);
		for (let file of files) {
			if (file.length == 0) continue;
			file = PATH.relative(PROCESS.cwd(), PATH.resolve(file));
			gitCurrentFiles.add(file);
		}
	}

	// Normalize file given to us.
	file = PATH.relative(PROCESS.cwd(), PATH.resolve(file));

	// Check and return if it's in here.
	return gitCurrentFiles.has(file);
}

async function git_retrieveAuthors(file) {
	console.debug(arguments.callee.name, Array.from(arguments));
	// git --no-pager log --date-order --reverse "--format=format:%aI|%aN <%aE>" -- file
	let lines = [];
	let result = await gitRL.queue(runProcessAsync, "git", ["--no-pager", "log", "--follow", `--format=format:${commitFormat}`, "--",	file], {});
	if (result[0] != 0) {
		let output = "";
		for (let message of result[1]) {
			output += message.text;
		}
		throw new Error(output);
	} else {
		let buffered = "";
		for (let message of result[1]) {
			if (message.type == "output")
				buffered += message.text;
		}
		lines = buffered.split(buffered.indexOf("\r\n") >= 0 ? "\r\n" : "\n");
	}

	if (await git_isInCurrentCommit(file)) {
		lines.push(await git_getCurrentAuthor());
	}

	let authors = new Map();
	for (let line of lines) {
		let [date, name] = line.split("|");

		let author = authors.get(name);
		if (author) {
			let dt = new Date(date)
			if (author.from > dt)
				author.from = dt;
			if (author.to < dt)
				author.to = dt;
		} else {
			authors.set(name, {
				from: new Date(date),
				to: new Date(date),
			})
		}
	}

	return authors;
}

async function generateCopyright(file) {
	console.debug(arguments.callee.name, Array.from(arguments));
	let authors = await git_retrieveAuthors(file)
	authors = new Map([...authors].sort((a, b) => {
		if (a[1].from < b[1].from) {
			return -1;
		} else if (a[1].from > b[1].from) {
			return 1;
		}
		return 0;
	}));
	let lines = [];
	for (let entry of authors) {
		let from = entry[1].from.getUTCFullYear();
		let to = entry[1].to.getUTCFullYear();
		lines.push(`Copyright (C) ${from != to ? `${from}-${to}` : to} ${entry[0]}`);
	}
	return lines;
}

function makeHeader(file, copyright) {
	console.debug(arguments.callee.name, Array.from(arguments));
	let file_name = PATH.basename(file).toLocaleLowerCase();
	let file_exts = file_name.substring(file_name.indexOf("."));

	let styles = formatStyleList;

	for (let key in styles) {
		let style = [key, styles[key]];
		if (style[1].files.includes(file_name)
			|| style[1].files.includes(file)
			|| style[1].exts.includes(file_exts)) {
			let header = [];
			header.push(...style[1].prepend);
			for (let line of copyright) {
				header.push(`${style[1].prefix}${line}${style[1].suffix}`);
			}
			header.push(...style[1].append);
			return header;
		}
	}

	throw new Error("Unrecognized file format.")
}

async function updateFile(file) {
	console.debug(arguments.callee.name, Array.from(arguments));
	await workRL.queue(async () => {
		try {
			if (abortAllWork) {
				return;
			}

			// Copyright information.
			let copyright = await generateCopyright(file);
			let header = undefined;
			try {
				header = makeHeader(file, copyright);
			} catch (ex) {
				console.log(`Skipping file '${file}': ${ex.message}`);
				return;
			}
			console.log(`Updating file '${file}'...`);

			// ToDo: Do we actually need to read the file first, or can we use the same rw stream?
			// File contents.
			let contentBuf = await FSPROMISES.readFile(file);
			let eol = contentBuf.indexOf("\r\n") != -1 ? "\r\n" : "\n";
			let headerBuf = Buffer.from(header.join(eol) + eol, "utf8");

			// Find the starting point.
			let startHeader = contentBuf.indexOf(Buffer.from(header[0], "utf8"));
			if (startHeader != -1) {
				//startHeader = contentBuf.lastIndexOf(eolBuf, startHeader);
				//startHeader += eolb.byteLength;
			}
			console.log(sectionStart, startHeader);

			// Find the ending point.
			let endHeader = contentBuf.lastIndexOf(Buffer.from(header[header.length - 1], "utf8"));
			if (endHeader != -1) {
				endHeader += Buffer.from(header[header.length - 1], "utf8").byteLength;
				endHeader += Buffer.byteLength(eol, "utf8");
			}
			console.log(sectionEnd, endHeader);

			// Last check for early-exit here.
			if (abortAllWork) {
				return;
			}
			
			let fd = await FSPROMISES.open(file, "w");
			if (startHeader == -1 || (endHeader < startHeader)) {
				fd.write(headerBuf, 0, null, 0);
				fd.write(contentBuf, 0, null, headerBuf.byteLength);
			} else {
				let pos = 0;

				if (startHeader > 0) {
					fd.write(contentBuf, 0, startHeader - Buffer.byteLength(eol, "utf8").byteLength, 0);
					pos += startHeader;
				}
				fd.write(headerBuf, 0, null, pos); pos += headerBuf.byteLength;
				fd.write(contentBuf, endHeader, null, pos);
			}
			await fd.close();
		} catch (ex) {
			console.error(`Error processing '${file}'!: ${ex}`);
			abortAllWork = true;
			PROCESS.exitCode = 1;
			return;
		}
	});
}

async function scanPath(path) {
	console.debug(arguments.callee.name, Array.from(arguments));

	// Abort here if the user aborted the process, or if the path is ignored.
	if (abortAllWork) {
		return;
	}

	console.log(`Scanning path '${path}'...`);

	let promises = [];

	await workRL.queue(async () => {
		let files = await FSPROMISES.readdir(path, { "withFileTypes": true });
		for (let file of files) {
			if (abortAllWork) {
				break;
			}

			let fullname = PATH.join(path, file.name);
			if (await git_isIgnored(fullname)) {
				console.log(`Ignoring path '${fullname}'...`);
				continue;
			}

			if (file.isDirectory()) {
				promises.push(scanPath(fullname));
			} else {
				promises.push(updateFile(fullname));
			}
		}
	});

	await Promise.all(promises);
}

(async function () {
	PROCESS.on("SIGINT", () => {
		abortAllWork = true;
		PROCESS.exitCode = 1;
		console.log("Sanely aborting all pending work...");
	})

	const root_path = PROCESS.cwd();
	let is_git_directory = false;
	
	console.debug(root_path, PROCESS.argv, PROCESS.execArgv);

	var args = PROCESS.argv.slice(2);
	let promises = new Array();
	while (args.length > 0) {
		// Try to place ourselves where git actually is.
		while (!is_git_directory) {
			if (abortAllWork) {
				return;
			}

			let entries = await FSPROMISES.readdir(PROCESS.cwd());
			if (entries.includes(".git") && ((await FSPROMISES.stat(PATH.join(PROCESS.cwd(), ".git"))).isDirectory())) {
				console.log(`Found .git at '${process.cwd()}'.`);
				is_git_directory = true;
			} else {
				PROCESS.chdir(PATH.resolve(PATH.join(PROCESS.cwd(), "..")));
			}
		}

		// If that failed, we just exit now.
		if (!is_git_directory) {
			console.error("Failed to figure out where git is, is this even a repository?");
			PROCESS.exitCode = 1;
			return;
		}

		if (abortAllWork) {
			return;
		}

		git_getCurrentAuthor();

		// Then proceed with normal work.
		let path = PATH.normalize(PATH.relative(process.cwd(), PATH.resolve(root_path, args[0])));

		// If we're handed a path, scan said path, otherwise update the file itself.
		let pathStat = await FSPROMISES.stat(path);
		if (pathStat) {
			if (await git_isIgnored(path)) {
				console.log(`Ignoring path '${path}'...`);
			} else if(pathStat.isDirectory()) {
				//await scanPath(path);
				promises.push(scanPath(path));
			} else {
				//await updateFile(path);
				promises.push(updateFile(path));
			}
		}

		// Slice off the first argument and continue.
		args = args.slice(1);
	}
	await Promise.all(promises);
	console.log("Done");
})();