BBB parser

src/BBBParser.cc

@@ -0,0 +1,269 @@
+#include "BBBParser.h"
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+
+int BBBParser::get() {
+	int ch = stream_.get();
+	if (ch == '\n') {
+		line_ += 1;
+		ch_ = 1;
+	} else {
+		ch_ += 1;
+	}
+
+	return ch;
+}
+
+void BBBParser::skip(char expected) {
+	int ch = get();
+	if (ch == EOF) {
+		error(std::string("Expected '") + expected + "', got EOF");
+	} else if (ch != expected) {
+		error(std::string("Expected '") + expected + "', got '" + (char)ch + "'");
+	}
+}
+
+void BBBParser::error(std::string msg) {
+	throw BBBParseError(std::to_string(line_) + ":" + std::to_string(ch_) + ": " + msg);
+}
+
+static bool isWhitespace(int ch) {
+	if (ch == ' ' || ch == '\t' || ch == '\r' || ch == '\n')
+		return true;
+	return false;
+}
+
+void BBBParser::skipWhitespaceLine() {
+	char ch;
+	while (isWhitespace(ch = peek()) && ch != '\r' && ch != '\n')
+		get();
+}
+
+void BBBParser::skipWhitespace() {
+	while (isWhitespace(peek()))
+		get();
+}
+
+char BBBParser::parseEscape() {
+	skip(); // '\'
+	int ch;
+	switch (ch = get()) {
+		case EOF:
+			error("Unexpected EOF");
+
+		case 'n':
+			return '\n';
+
+		case 'r':
+			return '\r';
+
+		case 't':
+			return '\t';
+
+		default:
+			return (char)ch;
+	}
+}
+
+void BBBParser::parseExpansion(const Variables &vars, std::vector<std::string> &values) {
+	skip(); // '$'
+
+	auto appendVar = [&](const std::string &name) {
+		if (name.size() == 0)
+			return;
+
+		auto it = vars.find(name);
+		if (it == vars.end())
+			return;
+
+		auto &vec = it->second;
+		for (auto &part: vec) {
+			values.push_back(part);
+		}
+	};
+
+	std::string str;
+	switch (peek()) {
+	case '{':
+		skip();
+		parseString(vars, str, '}');
+		skip('}');
+		appendVar(str);
+		break;
+
+	default:
+		if (!parseString(vars, str)) {
+			error("No identifier after $. Did you mean '\\$'?");
+		}
+
+		appendVar(str);
+	}
+}
+
+void BBBParser::parseQuotedExpansion(const Variables &vars, std::string &content) {
+	skip(); // '$'
+
+	auto appendVar = [&](const std::string &name) {
+		if (name.size() == 0)
+			return;
+
+		auto it = vars.find(name);
+		if (it == vars.end())
+			return;
+
+		auto &vec = it->second;
+		bool first = true;
+		for (auto &part: vec) {
+			if (!first) {
+				content += ' ';
+			}
+
+			first = false;
+			content += part;
+		}
+	};
+
+	std::string str;
+	switch (peek()) {
+	case '{':
+		skip();
+		parseString(vars, str, '}');
+		skip('}');
+		appendVar(str);
+		break;
+
+	default:
+		if (!parseString(vars, str, '"')) {
+			error("No identifier after $. Did you mean '\\$'?");
+		}
+
+		appendVar(str);
+	}
+}
+
+void BBBParser::parseQuotedString(const Variables &vars, std::string &content) {
+	skip(); // '"'
+
+	int ch;
+	while ((ch = peek()) != EOF) {
+		switch (ch) {
+		case EOF:
+			error("Unexpected EOF");
+
+		case '\\':
+			content.push_back(parseEscape());
+			break;
+
+		case '$':
+			parseQuotedExpansion(vars, content);
+			break;
+
+		case '"':
+			skip();
+			return;
+
+		default:
+			content.push_back(get());
+			break;
+		}
+	}
+}
+
+bool BBBParser::parseString(const Variables &vars, std::string &content, int sep) {
+	bool success = false;
+	int ch;
+	while (1) {
+		ch = peek();
+		if ((sep > 0 && ch == sep) || isWhitespace(ch)) {
+			return success;
+		}
+
+		switch (ch) {
+		case '=':
+		case EOF:
+			return success;
+
+		case '\\':
+			content.push_back(parseEscape());
+			break;
+
+		case '"':
+			parseQuotedString(vars, content);
+			success = true;
+			break;
+
+		default:
+			content.push_back(get());
+			success = true;
+			break;
+		}
+	}
+}
+
+void BBBParser::parseLine(const Variables &vars, std::vector<std::string> &values) {
+	std::string value;
+	while (true) {
+		skipWhitespaceLine();
+		if (peek() == EOF) {
+			break;
+		} else if (peek() == '$') {
+			parseExpansion(vars, values);
+			continue;
+		}
+
+		if (!parseString(vars, value)) {
+			break;
+		}
+
+		values.push_back(std::move(value));
+		value.clear();
+	}
+}
+
+void BBBParser::parse(Variables &vars) {
+	std::string key, value;
+	std::vector<std::string> values;
+
+	skipWhitespace();
+	if (!parseString(vars, key)) {
+		return;
+	}
+
+	skipWhitespace();
+	skip('=');
+
+	while (true) {
+		skipWhitespace();
+
+		if (peek() == EOF) {
+			break;
+		} else if (peek() == '$') {
+			parseExpansion(vars, values);
+			continue;
+		}
+
+		if (!parseString(vars, value)) {
+			break;
+		}
+
+		skipWhitespace();
+		if (peek() == '=') {
+			if (key.size() == 0) {
+				error("No variable on the left hand side of '='");
+			}
+
+			skip();
+			vars[key] = std::move(values);
+			key = std::move(value);
+			value.clear();
+		} else {
+			values.push_back(std::move(value));
+			value.clear();
+		}
+	}
+
+	vars[key] = std::move(values);
+}

src/BBBParser.h

@@ -0,0 +1,52 @@
+#pragma once
+
+#include <unordered_map>
+#include <vector>
+#include <string>
+#include <iostream>
+#include <exception>
+
+struct BBBParseError: std::exception {
+	BBBParseError(std::string msg): message(msg) {}
+	std::string message;
+
+	const char *what() const noexcept override {
+		return message.c_str();
+	}
+};
+
+class BBBParser {
+public:
+	using Variables = std::unordered_map<std::string, std::vector<std::string>>;
+
+	static const int FLAG_NONE = 0;
+	static const int FLAG_INSECURE = 1 << 0;
+
+	BBBParser(std::istream &stream, int flags, int line = 1, int ch = 1):
+		flags_(flags), line_(line), ch_(ch), stream_(stream) {}
+
+	void parseLine(const Variables &vars, std::vector<std::string> &values);
+	void parse(Variables &vars);
+
+	int line_;
+	int ch_;
+	int flags_;
+
+private:
+	int peek() { return stream_.peek(); }
+	int get();
+	void skip(char);
+	void skip() { get(); }
+	void error(std::string);
+
+	void skipWhitespaceLine();
+	void skipWhitespace();
+
+	char parseEscape();
+	void parseExpansion(const Variables &vars, std::vector<std::string> &values);
+	void parseQuotedExpansion(const Variables &vars, std::string &content);
+	void parseQuotedString(const Variables &vars, std::string &content);
+	bool parseString(const Variables &vars, std::string &content, int sep = -1);
+
+	std::istream &stream_;
+};