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(), O_RDONLY
#include <unistd.h>		// for close()
#include <errno.h>		// for perror()

#include <iostream>
#include <fstream>
#include <string>
#include <vector>
#include <map>
#include <set>
#include <functional>

#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) {
	iterator current = start;

	while(current != end && !limiter(current)) {
		++current;
	}

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


template <typename iterator>
std::string readBrackets(iterator &current, const iterator &end, const char * brackets) {
	if (current == end || *current != brackets[0]) {
		std::cout << brackets[0] << "!=" << *current;
		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;
}


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

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;
	}

	// create iteratable memory object
	MemoryString file((char*)memptr, fileinfo.st_size);

	std::string basedir = Path::Dir(path);

	// evaluate memory region (aka the file)
	list = (*this)(basedir, file);

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

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

// make a relative filepath absolute using the root filename
// and/or the process working directory
Path::Path fixFilename(const Path::Path &file, const Path::Path &root) {
	Path::Path temp(file);

	// try to make path absolute
	if (Path::isRelative(temp)) {
		// add current file's path
		temp = Path::Join(root, temp);
	}
	if (Path::isRelative(temp)) {
		// add process working directory
		temp = Path::Join(fs::cwd(), temp);
	}

	return Path::Clean(temp);
}

// evaluate a tex file searching for input statements
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";
		a = fixFilename(a, file);

		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"] = [file](List &l, std::string a){ if (!a.empty()) { l.push_back(fixFilename(a, file)); } };

	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;
}

struct Config {
	std::string targetName;
	Path::Path outfilePath;
	bool outfileOverride;
} config ;

std::ostream &writeTarget(std::ostream &stream, const std::string &target_name) {
	stream << target_name << ":";
	return stream;
}

std::ostream &writeTargetDepends(std::ostream &stream, const Path::Path &path) {
	stream << path << "\\\n";
	return stream;
}

int openOutfile(std::ofstream &stream) {
	cout << "writing dependency rules to " << config.outfilePath << endl;

	stream.close();
	stream.open(config.outfilePath);

	if (!stream) {
		cerr << "could not write to " << config.outfilePath << endl;
		return 0;
	}

	if (!config.targetName.empty()) {
		writeTarget(stream, config.targetName);
	}

	return 1;
}



#include <getopt.h>

struct option long_opts[] = {
	{"output", required_argument, NULL, 'o'},
	{"target", required_argument, NULL, 't'},
	{}, // terminator
};


void help(int argc, char ** args) {
	cout << args[0] << " [-o output] files..." << endl;
	cout << "--output" << endl;
	cout << "-o\tby default the program will create a depfile for every input" << endl;
	cout << "\tnaming it like the tex file with a .d extension. By giving the -o Option" << endl;
	cout << "\tthe output fill instead be put in this file exclusivly" << endl;
}

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

	int long_index = 0;
	int opt = 0;
	while((opt = getopt_long(argc, args, "o:t:", long_opts, &long_index)) != -1) {
		switch(opt) {
		case 'o':
			config.outfilePath = optarg;
			config.outfileOverride = true;
			break;
		case 't':
			config.targetName = optarg;
			break;
		case '?':
			help(argc, args);
			return 0;
		}
	}

	std::ofstream outfile;

	if (!config.outfilePath.empty() && !openOutfile(outfile)) {
		return -1; // failed to open outfile
	}

	// scan remaining arguments as input filenames
	for(; optind < argc; optind++) {
		Path::Path filename = args[optind];
		InputExtractor parser;

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

		// output results in makefile rule style

		// check for open outfile
		if (!config.outfileOverride) {
			config.outfilePath = Path::Basename(filename) + ".d";

			if (!openOutfile(outfile)) {
				continue; // just skip this file
			}

			writeTarget(outfile, filename);
		}

		for (auto it = list.begin(); it != list.end(); it++) {
			writeTargetDepends(outfile, *it);
		}

		// add newline for cleanness
		outfile << endl;

		cout << "done" << endl;
	}

}