texdepends/main.cpp

/*
 * main.cpp
 *
 *  Created on: 07.10.2017
 *      Author: julian
 */

#include <sys/mman.h> 	// for mmap()
#include <sys/stat.h>	// for fstat()
#include <fcntl.h>		// for open()
#include <unistd.h>		// for close()
#include <errno.h>		// for perror()

#include <iostream>
#include <string>
#include <cstring>
#include <vector>
#include <map>
#include <set>

#include "fs.hpp"
#include "memory_string.hpp"

using namespace std;

class InputExtractor
{
public:
	class Exception : public std::runtime_error {
	public:
		Exception(const std::string &str) : std::runtime_error(str) {}
	};

	typedef std::vector<Path::Path> List;

	List operator()(const Path::Path &file, const MemoryString &str);

	std::string macroExpand(const std::string &input);
	List Include(Path::Path);

protected:
	std::map<std::string, std::string> macros;
	std::set<Path::Path> includes;
};

#include <functional>


template <typename iterator>
std::string readTill(iterator &start, const iterator &end, std::function<bool(const iterator&)> limiter) {
	while(start != end && !limiter(start)) {
		++start;
	}

	return std::string(start, end);
}


template <typename iterator>
std::string readBrackets(iterator &begin, const iterator &end, const char * brackets) {
	auto current = begin;

	if (begin == end || *current != brackets[0]) {
		return std::string();
	}

	// skip first opening bracket
	current++;

	int depth = 1;
	auto bbegin = current, bend = current;

	while(depth > 0 && current != end) {
		if (*current== brackets[0]) {
			depth++;
		} else if (*current == brackets[1]) {
			depth--;
		} else {
			bend = ++current;
		}
	}

	// advance beyond last bracket
	if (current != end)
		current++;


	return std::string(bbegin, bend);
}

std::string InputExtractor::macroExpand(const std::string &input) {
	std::string result;
	std::map<std::string, std::string>::iterator lookup;
	//cout << "expanding: " << input << endl;

	std::string::const_iterator current = input.begin();
	while(current != input.end()) {
		if (*current == '\\') {
			current++;
			std::string::const_iterator start = current;

			while(current != input.end() && *current != '\\') {
				if ((lookup = macros.find(std::string(start, current))) != macros.end()) {
					break;
				}
			}

			if (lookup == macros.end()) {
				throw Exception("unknown macro in macro expansion: " + std::string(start, current));
			}

			result += lookup->second;
		} else {
			result += *current;
		}
		++current;
	}

	return result;
}

#include <functional>
typedef std::map<std::string, std::function<void(InputExtractor::List&, std::string)>> CommandList;

#include <unistd.h>


#include <fstream>

bool Exists(std::string path) {
	std::ifstream file(path);
	if (!file) {
		return false;
	}

	file.close();
	return true;
}

InputExtractor::List InputExtractor::Include(Path::Path path) {
	path = Path::Clean(path);
	List list;

	std::cout << "including file " << path << "...";

	// look for include
	if (includes.find(path) != includes.end()) {
		cout << "SKIP" << endl;
		return list; // already been there
	}

	cout << endl;

	// add to include list
	includes.insert(path);

	int fd = open(path.c_str(), O_RDONLY);
	if (fd == -1) {
		cerr << "cannot open " << path << endl;
		return list;
	}

	struct stat fileinfo;
	if (fstat(fd, &fileinfo) == -1) {
		perror("stat");
		close(fd);
		return list;
	}

	void * memptr = mmap(NULL, fileinfo.st_size, PROT_READ, MAP_SHARED, fd, 0);

	if (memptr == NULL) {
		perror("mmap");
		close(fd);
		return list;
	}

//	Substring str((const char *)memptr, (const char*)memptr + fileinfo.st_size);
	MemoryString file((char*)memptr, fileinfo.st_size);

	std::string basedir = Path::Dir(path);
	list = (*this)(basedir, file); // follow include

	// cleanup
	munmap(memptr, fileinfo.st_size);
	close(fd);
	return list;

	cout << path << "done" << endl;
}

InputExtractor::List InputExtractor::operator()(const Path::Path &file, const MemoryString &str){
	List result;
	CommandList IncludeCommands;

	IncludeCommands["input"] = [&file, this](List &l, std::string a) {
		if (a.empty()) return;
		if (Path::Extension(a) != ".tex") a += ".tex";
		// try to make path absolute
		if (Path::isRelative(a)) {
			// add current file's path
			a = Path::Join(file, a);
		}
		if (Path::isRelative(a)) {
			// add process working directory
			a = Path::Join(fs::cwd(), a);
		}

		l.push_back(a);
		// try to extract all inputs of that file
		auto sub = Include(a);
		if (!sub.empty())
			std::copy(sub.begin(), sub.end(), std::inserter(l, l.end()));
	};

	IncludeCommands["include"] = IncludeCommands["input"];
	IncludeCommands["lstinputlisting"] = [](List &l, std::string a){ l.push_back(a); };

	MemoryString::const_iterator current = str.begin();

	while(current != str.end()) {
		if (*current == '%') {
			// line commment
			while(current != str.end() && *current != '\n')
				current++;

			continue;
		}
		if (*current != '\\') {
			// skip non macros
			current++;
			continue;
		}

		// read macro name
		current++;

		if (current == str.end())
			throw Exception("unexpected EOF");

		auto start = current, end = current;

		auto limiter = [](char c) -> bool { return isspace(c) || c == '{' || c == '\\'; };

		for(; current != str.end() && !limiter(*current); end = (++current+1))
		{
			auto searchHit = IncludeCommands.find(std::string(start, end));
			if (searchHit != IncludeCommands.end()) {
				// handle crosslink
				current++;
				if (current == str.end()) continue;

				cout << searchHit->first << "[" << readBrackets(current, str.end(), "[]") << "]";
				auto inner = readBrackets(current, str.end(), "{}");
				cout << ":" << inner << endl;

				// add to results
				searchHit->second(result, macroExpand(inner));
				break;
			} else if (std::string(start, end) == std::string("def")) {
				// define new macro
				current++;
				if (current == str.end() || *current != '\\') {
					continue;
				}
				current++;

				std::function<bool(const MemoryString::const_iterator&)> limiter = [](const MemoryString::const_iterator &it) -> bool { return std::string("{ \t\n").find(*it) != std::string::npos; };

				std::string name = readTill(current, str.end(), limiter);

				cout << "new macro definition: " << name << endl;
				macros.insert(std::pair<std::string, std::string>(name, readBrackets(current, str.end(), "{}")));
				break;
			}
		}
	}

	return result;
}

int main(int argc, char ** args) {
	// find all the files the given tex files depend on
	int fd = 0;
	struct stat filestat;


	for(;argc > 1; --argc) {

		Path::Path filename = args[argc-1];
		InputExtractor parser;

		InputExtractor::List list = parser.Include(filename);

		// output results in makefile rule style

		Path::Path outfile_name = Path::Basename(filename) + ".d";
		cout << "writing dependecy rules to " << outfile_name << "...";

		std::ofstream outfile(outfile_name);

		if (!outfile) {
			cout << "could not create file!" << endl;
			continue;
		}

		outfile << filename << ":";

		for (auto it = list.begin(); it != list.end(); it++) {
			outfile << *it << "\t\\\n";
		}

		cout << "done" << endl;

		/*
		char * filename = args[argc-1];
		cout << "opening " << filename << "...";

		// try to open file
		fd = open(filename, O_RDONLY);
		if (fd == -1) {
			perror("could not open input file");
			continue;
		}

		fstat(fd, &filestat);
		//cout << "file size: " << filestat.st_size << endl;

		// try to mmap file
		void * memory_area = mmap(NULL, filestat.st_size, PROT_READ, MAP_SHARED, fd, 0);
		if (memory_area == nullptr) {
			perror("could not mmap the input");
			continue;
		}

		cout << "start parsing" << endl;

		MemoryString file((char*)memory_area, filestat.st_size);

		try {
			InputExtractor::List list = InputExtractor()(file);

			Path::Path outfilename = Path::Basename(Path::Path(filename)) + ".d";

			cout << "writing makedeps file to " << outfilename << "..." << endl;

			// write in makefile style
			std::ofstream output(outfilename);
			if (!output) {
				std::cout << "could not create output file" << std::endl;
			} else {
				output << filename << ": ";

				for (auto it = list.begin(); it != list.end(); it++) {
					if (Path::isRelative(*it))  {
						if (Path::isRelative(filename)) {
							*it = Path::Join(fs::cwd(), filename, *it);
						} else {
							*it = Path::Join(Path::Path(filename), *it);
						}
					}

					cout << "depends: " << *it << endl;
					output << '\t' << *it << "\t\\\n";
				}
				output << endl;
			}

			cout << filename << done;

		} catch(InputExtractor::Exception &e) {
			cout << e.what() << endl;
		}

		// cleanup
		munmap(memory_area, filestat.st_size);
		close(fd);

		*/
	}

}