#include <utils.h>
#include <strprintf.h>
#include <logger.h>
#include <config.h>

#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <iconv.h>
#include <errno.h>
#include <pwd.h>
#include <libgen.h>
#include <sys/utsname.h>
#include <unistd.h>
#include <sys/param.h>

#include <unordered_set>
#include <unistd.h>
#include <sstream>
#include <locale>
#include <cwchar>
#include <cstring>
#include <cstdlib>
#include <cstdarg>
#include <cstdio>
#include <algorithm>
#include <regex>

#include <curl/curl.h>

#include <langinfo.h>
#include <stfl.h>
#include <libxml/uri.h>

#if HAVE_GCRYPT
#include <gnutls/gnutls.h>
#include <gcrypt.h>
#include <errno.h>
#include <pthread.h>
GCRY_THREAD_OPTION_PTHREAD_IMPL;
#endif

#if HAVE_OPENSSL && OPENSSL_VERSION_NUMBER < 0x01010000fL
#include <openssl/crypto.h>
#endif

namespace newsbeuter {

std::vector<std::string> utils::tokenize_quoted(const std::string& str, std::string delimiters) {
	/*
	 * This function tokenizes strings, obeying quotes and throwing away comments that start
	 * with a '#'.
	 *
	 * e.g. line: foo bar "foo bar" "a test"
	 * is parsed to 4 elements:
	 * 	[0]: foo
	 * 	[1]: bar
	 * 	[2]: foo bar
	 * 	[3]: a test
	 *
	 * e.g. line: yes great "x\ny" # comment
	 * is parsed to 3 elements:
	 * 	[0]: yes
	 * 	[1]: great
	 * 	[2]: x
	 * 	y
	 *
	 * 	\", \r, \n, \t and \v are replaced with the literals that you know from C/C++ strings.
	 *
	 */
	bool attach_backslash = true;
	std::vector<std::string> tokens;
	std::string::size_type last_pos = str.find_first_not_of(delimiters, 0);
	std::string::size_type pos = last_pos;

	while (pos != std::string::npos && last_pos != std::string::npos) {
		if (str[last_pos] == '#') // stop as soon as we found a comment
			break;

		if (str[last_pos] == '"') {
			++last_pos;
			pos = last_pos;
			int backslash_count = 0;
			while (pos < str.length() && (str[pos] != '"' || (backslash_count%2))) {
				if (str[pos] == '\\') {
					++backslash_count;
				} else {
					backslash_count = 0;
				}
				++pos;
			}
			if (pos >= str.length()) {
				pos = std::string::npos;
				std::string token;
				while (last_pos < str.length()) {
					if (str[last_pos] == '\\') {
						if (str[last_pos-1] == '\\') {
							if (attach_backslash) {
								token.append("\\");
							}
							attach_backslash = !attach_backslash;
						}
					} else {
						if (str[last_pos-1] == '\\') {
							append_escapes(token, str[last_pos]);
						} else {
							token.append(1, str[last_pos]);
						}
					}
					++last_pos;
				}
				tokens.push_back(token);
			} else {
				std::string token;
				while (last_pos < pos) {
					if (str[last_pos] == '\\') {
						if (str[last_pos-1] == '\\') {
							if (attach_backslash) {
								token.append("\\");
							}
							attach_backslash = !attach_backslash;
						}
					} else {
						if (str[last_pos-1] == '\\') {
							append_escapes(token, str[last_pos]);
						} else {
							token.append(1, str[last_pos]);
						}
					}
					++last_pos;
				}
				tokens.push_back(token);
				++pos;
			}
		} else {
			pos = str.find_first_of(delimiters, last_pos);
			tokens.push_back(str.substr(last_pos, pos - last_pos));
		}
		last_pos = str.find_first_not_of(delimiters, pos);
	}

	return tokens;
}


std::vector<std::string> utils::tokenize(const std::string& str, std::string delimiters) {
	/*
	 * This function tokenizes a string by the delimiters. Plain and simple.
	 */
	std::vector<std::string> tokens;
	std::string::size_type last_pos = str.find_first_not_of(delimiters, 0);
	std::string::size_type pos = str.find_first_of(delimiters, last_pos);

	while (std::string::npos != pos || std::string::npos != last_pos) {
		tokens.push_back(str.substr(last_pos, pos - last_pos));
		last_pos = str.find_first_not_of(delimiters, pos);
		pos = str.find_first_of(delimiters, last_pos);
	}
	return tokens;
}

std::vector<std::wstring> utils::wtokenize(const std::wstring& str, std::wstring delimiters) {
	/*
	 * This function tokenizes a string by the delimiters. Plain and simple.
	 */
	std::vector<std::wstring> tokens;
	std::wstring::size_type last_pos = str.find_first_not_of(delimiters, 0);
	std::wstring::size_type pos = str.find_first_of(delimiters, last_pos);

	while (std::string::npos != pos || std::string::npos != last_pos) {
		tokens.push_back(str.substr(last_pos, pos - last_pos));
		last_pos = str.find_first_not_of(delimiters, pos);
		pos = str.find_first_of(delimiters, last_pos);
	}
	return tokens;
}

std::vector<std::string> utils::tokenize_spaced(const std::string& str, std::string delimiters) {
	std::vector<std::string> tokens;
	std::string::size_type last_pos = str.find_first_not_of(delimiters, 0);
	std::string::size_type pos = str.find_first_of(delimiters, last_pos);

	if (last_pos != 0) {
		tokens.push_back(str.substr(0, last_pos));
	}

	while (std::string::npos != pos || std::string::npos != last_pos) {
		tokens.push_back(str.substr(last_pos, pos - last_pos));
		last_pos = str.find_first_not_of(delimiters, pos);
		if (last_pos > pos)
			tokens.push_back(str.substr(pos, last_pos - pos));
		pos = str.find_first_of(delimiters, last_pos);
	}

	return tokens;
}

std::string utils::consolidate_whitespace(const std::string& str, std::string whitespace) {
	std::string result;
	std::string::size_type last_pos = str.find_first_not_of(whitespace);
	std::string::size_type pos = str.find_first_of(whitespace, last_pos);

	if (last_pos != 0 && str != "") {
		result.append(str.substr(0, last_pos));
	}

	while (std::string::npos != pos || std::string::npos != last_pos) {
		result.append(str.substr(last_pos, pos - last_pos));
		last_pos = str.find_first_not_of(whitespace, pos);
		if (last_pos > pos)
			result.append(" ");
		pos = str.find_first_of(whitespace, last_pos);
	}

	return result;
}

std::vector<std::string> utils::tokenize_nl(const std::string& str, std::string delimiters) {
	std::vector<std::string> tokens;
	std::string::size_type last_pos = str.find_first_not_of(delimiters, 0);
	std::string::size_type pos = str.find_first_of(delimiters, last_pos);
	unsigned int i;

	LOG(level::DEBUG,"utils::tokenize_nl: last_pos = %u",last_pos);
	if (last_pos != std::string::npos) {
		for (i=0; i<last_pos; ++i) {
			tokens.push_back(std::string("\n"));
		}
	}

	while (std::string::npos != pos || std::string::npos != last_pos) {
		tokens.push_back(str.substr(last_pos, pos - last_pos));
		LOG(level::DEBUG,"utils::tokenize_nl: substr = %s", str.substr(last_pos, pos - last_pos));
		last_pos = str.find_first_not_of(delimiters, pos);
		LOG(level::DEBUG,"utils::tokenize_nl: pos - last_pos = %u", last_pos - pos);
		for (i=0; last_pos != std::string::npos && pos != std::string::npos && i<(last_pos - pos); ++i) {
			tokens.push_back(std::string("\n"));
		}
		pos = str.find_first_of(delimiters, last_pos);
	}

	return tokens;
}

void utils::remove_fs_lock(const std::string& lock_file) {
	LOG(level::DEBUG, "utils::remove_fs_lock: removed lockfile %s", lock_file);
	::unlink(lock_file.c_str());
}

bool utils::try_fs_lock(const std::string& lock_file, pid_t & pid) {
	int fd;
	// pid == 0 indicates that something went majorly wrong during locking
	pid = 0;

	LOG(level::DEBUG, "utils::try_fs_lock: trying to lock %s", lock_file);

	// first, we open (and possibly create) the lock file
	fd = ::open(lock_file.c_str(), O_RDWR | O_CREAT, 0600);
	if (fd < 0)
		return false;

	// then we lock it (T_LOCK returns immediately if locking is not possible)
	if (lockf(fd, F_TLOCK, 0) == 0) {
		std::string pidtext = std::to_string(getpid());
		// locking successful -> truncate file and write own PID into it
		ssize_t written = 0;
		if (ftruncate(fd, 0) == 0) {
			written = write(fd, pidtext.c_str(), pidtext.length());
		}
		bool success =
			   (written != -1)
			&& (static_cast<unsigned int>(written) == pidtext.length());
		return success;
	}

	// locking was not successful -> read PID of locking process from it
	fd = ::open(lock_file.c_str(), O_RDONLY);
	if (fd >= 0) {
		char buf[32];
		int len = read(fd, buf, sizeof(buf)-1);
		unsigned int upid = 0;
		buf[len] = '\0';
		sscanf(buf, "%u", &upid);
		pid = upid;
		close(fd);
	}
	return false;
}

std::string utils::translit(const std::string& tocode, const std::string& fromcode)
{
	std::string tlit = "//TRANSLIT";

	enum class translit_state {
		UNKNOWN,
		SUPPORTED,
		UNSUPPORTED
	};

	static translit_state state = translit_state::UNKNOWN;

	// TRANSLIT is not needed when converting to unicode encodings
	if (tocode == "utf-8" || tocode == "WCHAR_T") return tocode;

	if (state == translit_state::UNKNOWN) {
		iconv_t cd = ::iconv_open((tocode + "//TRANSLIT").c_str(), fromcode.c_str());

		if (cd == reinterpret_cast<iconv_t>(-1)) {
			if (errno == EINVAL) {
				iconv_t cd = ::iconv_open(tocode.c_str(), fromcode.c_str());
				if (cd != reinterpret_cast<iconv_t>(-1)) {
					state = translit_state::UNSUPPORTED;
				} else {
					fprintf(stderr, "iconv_open('%s', '%s') failed: %s", tocode.c_str(), fromcode.c_str(), strerror(errno));
					abort();
				}
			} else {
				fprintf(stderr, "iconv_open('%s//TRANSLIT', '%s') failed: %s", tocode.c_str(), fromcode.c_str(), strerror(errno));
				abort();
			}
		} else {
			state = translit_state::SUPPORTED;
		}

		iconv_close(cd);
	}

	return ((state == translit_state::SUPPORTED) ? (tocode + tlit) : (tocode));
}

std::string utils::convert_text(const std::string& text, const std::string& tocode, const std::string& fromcode) {
	std::string result;

	if (strcasecmp(tocode.c_str(), fromcode.c_str())==0)
		return text;

	iconv_t cd = ::iconv_open(translit(tocode, fromcode).c_str(), fromcode.c_str());

	if (cd == reinterpret_cast<iconv_t>(-1))
		return result;

	size_t inbytesleft;
	size_t outbytesleft;

	/*
	 * of all the Unix-like systems around there, only Linux/glibc seems to
	 * come with a SuSv3-conforming iconv implementation.
	 */
#if !(__linux) && !defined(__GLIBC__) && !defined(__APPLE__) \
	&& !defined(__OpenBSD__) && !defined(__FreeBSD__) && !defined(__DragonFly__)
	const char * inbufp;
#else
	char * inbufp;
#endif
	char outbuf[16];
	char * outbufp = outbuf;

	outbytesleft = sizeof(outbuf);
	inbufp = const_cast<char *>(text.c_str()); // evil, but spares us some trouble
	inbytesleft = strlen(inbufp);

	do {
		char * old_outbufp = outbufp;
		int rc = ::iconv(cd, &inbufp, &inbytesleft, &outbufp, &outbytesleft);
		if (-1 == rc) {
			switch (errno) {
			case E2BIG:
				result.append(old_outbufp, outbufp - old_outbufp);
				outbufp = outbuf;
				outbytesleft = sizeof(outbuf);
				inbufp += strlen(inbufp) - inbytesleft;
				inbytesleft = strlen(inbufp);
				break;
			case EILSEQ:
			case EINVAL:
				result.append(old_outbufp, outbufp - old_outbufp);
				result.append("?");
				inbufp += strlen(inbufp) - inbytesleft + 1;
				inbytesleft = strlen(inbufp);
				break;
			default:
				break;
			}
		} else {
			result.append(old_outbufp, outbufp - old_outbufp);
		}
	} while (inbytesleft > 0);

	iconv_close(cd);

	return result;
}

std::string utils::get_command_output(const std::string& cmd) {
	FILE * f = popen(cmd.c_str(), "r");
	std::string buf;
	if (f) {
		char cbuf[1024];
		size_t s;
		while ((s = fread(cbuf, 1, sizeof(cbuf), f)) > 0) {
			buf.append(cbuf, s);
		}
		pclose(f);
	}
	return buf;
}

void utils::extract_filter(const std::string& line, std::string& filter, std::string& url) {
	std::string::size_type pos = line.find_first_of(":", 0);
	std::string::size_type pos1 = line.find_first_of(":", pos + 1);
	filter = line.substr(pos+1, pos1 - pos - 1);
	pos = pos1;
	url = line.substr(pos+1, line.length() - pos);
	LOG(level::DEBUG, "utils::extract_filter: %s -> filter: %s url: %s", line, filter, url);
}

static size_t my_write_data(void *buffer, size_t size, size_t nmemb, void *userp) {
	std::string * pbuf = static_cast<std::string *>(userp);
	pbuf->append(static_cast<const char *>(buffer), size * nmemb);
	return size * nmemb;
}

std::string utils::retrieve_url(
		const std::string& url,
		configcontainer * cfgcont,
		const std::string& authinfo,
		const std::string* postdata)
{
	std::string buf;

	CURL * easyhandle = curl_easy_init();
	set_common_curl_options(easyhandle, cfgcont);
	curl_easy_setopt(easyhandle, CURLOPT_URL, url.c_str());
	curl_easy_setopt(easyhandle, CURLOPT_WRITEFUNCTION, my_write_data);
	curl_easy_setopt(easyhandle, CURLOPT_WRITEDATA, &buf);

	if (postdata != nullptr) {
		curl_easy_setopt(easyhandle, CURLOPT_POST, 1);
		curl_easy_setopt(easyhandle, CURLOPT_POSTFIELDS, postdata->c_str());
	}

	if (! authinfo.empty()) {
		curl_easy_setopt(easyhandle, CURLOPT_HTTPAUTH,
				get_auth_method(cfgcont->get_configvalue("http-auth-method")));
		curl_easy_setopt(easyhandle, CURLOPT_USERPWD, authinfo.c_str());
	}

	curl_easy_perform(easyhandle);
	curl_easy_cleanup(easyhandle);

	if (postdata != nullptr) {
		LOG(level::DEBUG, "utils::retrieve_url(%s)[%s]: %s", url, postdata, buf);
	} else {
		LOG(level::DEBUG, "utils::retrieve_url(%s)[-]: %s", url, buf);
	}

	return buf;
}

void utils::run_command(const std::string& cmd, const std::string& input) {
	int rc = fork();
	switch (rc) {
	case -1:
		break;
	case 0: { // child:
		int fd = ::open("/dev/null", O_RDWR);
		close(0);
		close(1);
		close(2);
		dup2(fd, 0);
		dup2(fd, 1);
		dup2(fd, 2);
		LOG(level::DEBUG, "utils::run_command: %s '%s'", cmd, input);
		execlp(cmd.c_str(), cmd.c_str(), input.c_str(), nullptr);
		LOG(level::DEBUG, "utils::run_command: execlp of %s failed: %s", cmd, strerror(errno));
		exit(1);
	}
	default:
		break;
	}
}

std::string utils::run_program(char * argv[], const std::string& input) {
	std::string buf;
	int ipipe[2];
	int opipe[2];
	if (pipe(ipipe) != 0) {
		return "";
	}
	if (pipe(opipe) != 0) {
		return "";
	}

	int rc = fork();
	switch (rc) {
	case -1:
		break;
	case 0: { // child:
		close(ipipe[1]);
		close(opipe[0]);
		dup2(ipipe[0], 0);
		dup2(opipe[1], 1);
		close(2);

		int errfd = ::open("/dev/null", O_WRONLY);
		if (errfd != -1) dup2(errfd, 2);

		execvp(argv[0], argv);
		exit(1);
	}
	default: {
		close(ipipe[0]);
		close(opipe[1]);
		ssize_t written = 0;
		written = write(ipipe[1], input.c_str(), input.length());
		if (written != -1) {
			close(ipipe[1]);
			char cbuf[1024];
			int rc2;
			while ((rc2 = read(opipe[0], cbuf, sizeof(cbuf))) > 0) {
				buf.append(cbuf, rc2);
			}
		} else {
			close(ipipe[1]);
		}
		close(opipe[0]);
	}
	break;
	}
	return buf;
}

std::string utils::resolve_tilde(const std::string& str) {
	const char * homedir;
	std::string filepath;

	if (!(homedir = ::getenv("HOME"))) {
		struct passwd * spw = ::getpwuid(::getuid());
		if (spw) {
			homedir = spw->pw_dir;
		} else {
			homedir = "";
		}
	}

	if (strcmp(homedir,"")!=0) {
		if (str == "~") {
			filepath.append(homedir);
		} else if (str.substr(0,2) == "~/") {
			filepath.append(homedir);
			filepath.append(1,'/');
			filepath.append(str.substr(2,str.length()-2));
		} else {
			filepath.append(str);
		}
	} else {
		filepath.append(str);
	}

	return filepath;
}

std::string utils::replace_all(std::string str, const std::string& from, const std::string& to) {
	std::string::size_type s = str.find(from);
	while (s != std::string::npos) {
		str.replace(s,from.length(), to);
		s = str.find(from, s + to.length());
	}
	return str;
}

std::wstring utils::utf8str2wstr(const std::string& utf8str) {
	stfl_ipool * pool = stfl_ipool_create("utf-8");
	std::wstring wstr = stfl_ipool_towc(pool, utf8str.c_str());
	stfl_ipool_destroy(pool);
	return wstr;
}

std::wstring utils::str2wstr(const std::string& str) {
	const char * codeset = nl_langinfo(CODESET);
	struct stfl_ipool * ipool = stfl_ipool_create(codeset);
	std::wstring result = stfl_ipool_towc(ipool, str.c_str());
	stfl_ipool_destroy(ipool);
	return result;
}

std::string utils::wstr2str(const std::wstring& wstr) {
	std::string codeset = nl_langinfo(CODESET);
	codeset = translit(codeset, "WCHAR_T");
	struct stfl_ipool * ipool = stfl_ipool_create(codeset.c_str());
	std::string result = stfl_ipool_fromwc(ipool, wstr.c_str());
	stfl_ipool_destroy(ipool);
	return result;
}

std::string utils::absolute_url(const std::string& url, const std::string& link) {
	xmlChar * newurl = xmlBuildURI((const xmlChar *)link.c_str(), (const xmlChar *)url.c_str());
	std::string retval;
	if (newurl) {
		retval = (const char *)newurl;
		xmlFree(newurl);
	} else {
		retval = link;
	}
	return retval;
}

std::string utils::get_useragent(configcontainer * cfgcont) {
	std::string ua_pref = cfgcont->get_configvalue("user-agent");
	if (ua_pref.length() == 0) {
		struct utsname buf;
		uname(&buf);
		if (strcmp(buf.sysname, "Darwin") == 0) {
			/* Assume it is a Mac from the last decade or at least Mac-like */
			const char* PROCESSOR = "";
			if (strcmp(buf.machine, "x86_64") == 0 || strcmp(buf.machine, "i386") == 0) {
				PROCESSOR = "Intel ";
			}
			return strprintf::fmt("%s/%s (Macintosh; %sMac OS X)", PROGRAM_NAME, PROGRAM_VERSION, PROCESSOR);
		}
		return strprintf::fmt("%s/%s (%s %s)", PROGRAM_NAME, PROGRAM_VERSION, buf.sysname, buf.machine);
	}
	return ua_pref;
}

unsigned int utils::to_u(
		const std::string& str,
		const unsigned int default_value)
{
	std::istringstream is(str);
	unsigned int u;
	is >> u;
	if (is.fail() || ! is.eof()) {
		u = default_value;
	}
	return u;
}

scope_measure::scope_measure(const std::string& func, level ll)
	: funcname(func), lvl(ll)
{
	gettimeofday(&tv1, nullptr);
}

void scope_measure::stopover(const std::string& son) {
	gettimeofday(&tv2, nullptr);
	unsigned long diff = (((tv2.tv_sec - tv1.tv_sec) * 1000000) + tv2.tv_usec) - tv1.tv_usec;
	LOG(lvl, "scope_measure: function `%s' (stop over `%s') took %lu.%06lu s so far", funcname, son, diff / 1000000, diff % 1000000);
}

scope_measure::~scope_measure() {
	gettimeofday(&tv2, nullptr);
	unsigned long diff = (((tv2.tv_sec - tv1.tv_sec) * 1000000) + tv2.tv_usec) - tv1.tv_usec;
	LOG(level::INFO, "scope_measure: function `%s' took %lu.%06lu s", funcname, diff / 1000000, diff % 1000000);
}

void utils::append_escapes(std::string& str, char c) {
	switch (c) {
	case 'n':
		str.append("\n");
		break;
	case 'r':
		str.append("\r");
		break;
	case 't':
		str.append("\t");
		break;
	case '"':
		str.append("\"");
		break;
	// escaped backticks are passed through, still escaped. We un-escape them
	// in configparser::evaluate_backticks
	case '`':
		str.append("\\`");
		break;
	case '\\':
		break;
	default:
		str.append(1, c);
		break;
	}
}

bool utils::is_valid_color(const std::string& color) {
	static const std::unordered_set<std::string> colors = {
		"black", "red", "green", "yellow", "blue",
		"magenta", "cyan", "white", "default"
	};
	if (colors.find(color) != colors.end()) {
		return true;
	}

	// does it start with "color"?
	if (color.size() > 5 && color.substr(0, 5) == "color") {
		if (color[5] == '0') {
			// if the remainder of the string starts with zero, it can only be
			// "color0"
			return color == "color0" ? true : false;
		} else {
			// we're now sure that the remainder doesn't start with zero, but
			// is it a valid decimal number?
			const std::string number = color.substr(5, color.size()-5);
			size_t pos {};
			int n = std::stoi(number, &pos);

			// remainder should not contain any trailing characters
			if (number.size() != pos) return false;

			// remainder should be a number in (0; 255]. The interval is
			// half-open because zero is already handled above.
			if (n > 0 && n < 256) return true;
		}
	}
	return false;
}

bool utils::is_valid_attribute(const std::string& attrib) {
	static const std::unordered_set<std::string> attribs = {
		"standout", "underline", "reverse", "blink",
		"dim", "bold", "protect", "invis", "default"
	};
	if (attribs.find(attrib) != attribs.end()) {
		return true;
	} else {
		return false;
	}
}

std::vector<std::pair<unsigned int, unsigned int>> utils::partition_indexes(unsigned int start, unsigned int end, unsigned int parts) {
	std::vector<std::pair<unsigned int, unsigned int>> partitions;
	unsigned int count = end - start + 1;
	unsigned int size = count / parts;

	for (unsigned int i=0; i<parts-1; i++) {
		partitions.push_back(std::pair<unsigned int, unsigned int>(start, start + size - 1));
		start += size;
	}

	partitions.push_back(std::pair<unsigned int, unsigned int>(start, end));
	return partitions;
}

size_t utils::strwidth(const std::string& str) {
	std::wstring wstr = str2wstr(str);
	int width = wcswidth(wstr.c_str(), wstr.length());
	if (width < 1) // a non-printable character found?
		return wstr.length(); // return a sane width (which might be larger than necessary)
	return width; // exact width
}

size_t utils::strwidth_stfl(const std::string& str) {
	size_t reduce_count = 0;
	size_t len = str.length();
	if (len > 1) {
		for (size_t idx=0; idx<len-1; ++idx) {
			if (str[idx] == '<' && str[idx+1] != '>') {
				reduce_count += 3;
				idx += 3;
			}
		}
	}

	return strwidth(str) - reduce_count;
}

size_t utils::wcswidth_stfl(const std::wstring& str, size_t size) {
	size_t reduce_count = 0;
	size_t len = std::min(str.length(), size);
	if (len > 1) {
		for (size_t idx=0; idx<len-1; ++idx) {
			if (str[idx] == L'<' && str[idx+1] != L'>') {
				reduce_count += 3;
				idx += 3;
			}
		}
	}

	int width = wcswidth(str.c_str(), size);
	if (width < 0) {
		LOG(level::ERROR, "oh, oh, wcswidth just failed");
		return str.length() - reduce_count;
	}

	return width - reduce_count;
}

std::string utils::join(const std::vector<std::string>& strings, const std::string& separator) {
	std::string result;

	for (auto str : strings) {
		result.append(str);
		result.append(separator);
	}

	if (result.length() > 0)
		result.erase(result.length()-separator.length(), result.length());

	return result;
}

bool utils::is_special_url(const std::string& url) {
	return url.substr(0,6) == "query:" || url.substr(0,7) == "filter:" || url.substr(0,5) == "exec:";
}

bool utils::is_http_url(const std::string& url) {
	return url.substr(0,7) == "http://" || url.substr(0,8) == "https://";
}

std::string utils::censor_url(const std::string& url) {
	std::string rv;
	if (url.length() > 0 && !utils::is_special_url(url)) {
		const char * myuri = url.c_str();
		xmlURIPtr uri = xmlParseURI(myuri);
		if (uri) {
			if (uri->user) {
				xmlFree(uri->user);
				uri->user = (char *)xmlStrdup((const xmlChar *)"*:*");
			}
			xmlChar * uristr = xmlSaveUri(uri);

			rv = (const char *)uristr;
			xmlFree(uristr);
			xmlFreeURI(uri);
		} else
			return url;
	} else {
		rv = url;
	}
	return rv;
}

std::string utils::quote_for_stfl(std::string str) {
	unsigned int len = str.length();
	for (unsigned int i=0; i<len; ++i) {
		if (str[i] == '<') {
			str.insert(i+1, ">");
			++len;
		}
	}
	return str;
}

void utils::trim(std::string& str) {
	while (str.length() > 0 && ::isspace(str[0])) {
		str.erase(0,1);
	}
	trim_end(str);
}

void utils::trim_end(std::string& str) {
	std::string::size_type pos = str.length()-1;
	while (str.length()>0 && (str[pos] == '\n' || str[pos] == '\r')) {
		str.erase(pos);
		pos--;
	}
}

std::string utils::quote(const std::string& str) {
	std::string rv = replace_all(str, "\"", "\\\"");
	rv.insert(0, "\"");
	rv.append("\"");
	return rv;
}

unsigned int utils::get_random_value(unsigned int max) {
	static bool initialized = false;
	if (!initialized) {
		initialized = true;
		srand(~(time(nullptr) ^ getpid() ^ getppid()));
	}
	// OpenBSD will warn you that rand() can be deterministic. We don't care,
	// because this function is only used for simple things like selecting
	// a random unread item.
	return static_cast<unsigned int>(rand() % max);
}

std::string utils::quote_if_necessary(const std::string& str) {
	std::string result;
	if (str.find_first_of(" ", 0) == std::string::npos) {
		result = str;
	} else {
		result = utils::replace_all(str, "\"", "\\\"");
		result.insert(0, "\"");
		result.append("\"");
	}
	return result;
}

void utils::set_common_curl_options(CURL * handle, configcontainer * cfg) {
	std::string proxy;
	std::string proxyauth;
	std::string proxyauthmethod;
	std::string proxytype;
	std::string useragent;
	std::string cookie_cache;
	unsigned int dl_timeout = 0;

	if (cfg) {
		if (cfg->get_configvalue_as_bool("use-proxy")) {
			proxy = cfg->get_configvalue("proxy");
			proxyauth = cfg->get_configvalue("proxy-auth");
			proxyauthmethod = cfg->get_configvalue("proxy-auth-method");
			proxytype = cfg->get_configvalue("proxy-type");
		}
		useragent = utils::get_useragent(cfg);
		dl_timeout = cfg->get_configvalue_as_int("download-timeout");
		cookie_cache = cfg->get_configvalue("cookie-cache");
	}

	curl_easy_setopt(handle, CURLOPT_SSL_VERIFYHOST, cfg->get_configvalue_as_bool("ssl-verifyhost") ? 2 : 0);
	curl_easy_setopt(handle, CURLOPT_SSL_VERIFYPEER, cfg->get_configvalue_as_bool("ssl-verifypeer"));
	curl_easy_setopt(handle, CURLOPT_NOSIGNAL, 1);
	curl_easy_setopt(handle, CURLOPT_ENCODING, "gzip, deflate");
	curl_easy_setopt(handle, CURLOPT_TIMEOUT, dl_timeout);

	if (proxy != "")
		curl_easy_setopt(handle, CURLOPT_PROXY, proxy.c_str());
	if (proxyauth != "") {
		curl_easy_setopt(handle, CURLOPT_PROXYAUTH, get_auth_method(proxyauthmethod));
		curl_easy_setopt(handle, CURLOPT_PROXYUSERPWD, proxyauth.c_str());
	}
	if (proxytype != "") {
		LOG(level::DEBUG, "utils::set_common_curl_options: proxytype = %s", proxytype);
		curl_easy_setopt(handle, CURLOPT_PROXYTYPE, get_proxy_type(proxytype));
	}

	curl_easy_setopt(handle, CURLOPT_USERAGENT, useragent.c_str());

	curl_easy_setopt(handle, CURLOPT_FOLLOWLOCATION, 1);
	curl_easy_setopt(handle, CURLOPT_MAXREDIRS, 10);
	curl_easy_setopt(handle, CURLOPT_FAILONERROR, 1);

	if (cookie_cache != "") {
		curl_easy_setopt(handle, CURLOPT_COOKIEFILE, cookie_cache.c_str());
		curl_easy_setopt(handle, CURLOPT_COOKIEJAR, cookie_cache.c_str());
	}
}

std::string utils::get_content(xmlNode * node) {
	std::string retval;
	if (node) {
		xmlChar * content = xmlNodeGetContent(node);
		if (content) {
			retval = (const char *)content;
			xmlFree(content);
		}
	}
	return retval;
}

unsigned long utils::get_auth_method(const std::string& type) {
	if (type == "any")
		return CURLAUTH_ANY;
	if (type == "basic")
		return CURLAUTH_BASIC;
	if (type == "digest")
		return CURLAUTH_DIGEST;
#ifdef CURLAUTH_DIGEST_IE
	if (type == "digest_ie")
		return CURLAUTH_DIGEST_IE;
#else
# warning "proxy-auth-method digest_ie not added due to libcurl older than 7.19.3"
#endif
	if (type == "gssnegotiate")
		return CURLAUTH_GSSNEGOTIATE;
	if (type == "ntlm")
		return CURLAUTH_NTLM;
	if (type == "anysafe")
		return CURLAUTH_ANYSAFE;
	if (type != "") {
		LOG(level::USERERROR, "you configured an invalid proxy authentication method: %s", type);
	}
	return CURLAUTH_ANY;
}

curl_proxytype utils::get_proxy_type(const std::string& type) {
	if (type == "http")
		return CURLPROXY_HTTP;
	if (type == "socks4")
		return CURLPROXY_SOCKS4;
	if (type == "socks5")
		return CURLPROXY_SOCKS5;
	if (type == "socks5h")
		return CURLPROXY_SOCKS5_HOSTNAME;
#ifdef CURLPROXY_SOCKS4A
	if (type == "socks4a")
		return CURLPROXY_SOCKS4A;
#endif

	if (type != "") {
		LOG(level::USERERROR, "you configured an invalid proxy type: %s", type);
	}
	return CURLPROXY_HTTP;
}

std::string utils::escape_url(const std::string& url) {
	return replace_all(replace_all(url,"?","%3F"), "&", "%26");
}

std::string utils::unescape_url(const std::string& url) {
	return replace_all(replace_all(url,"%3F","?"), "%26", "&");
}

std::wstring utils::clean_nonprintable_characters(std::wstring text) {
	for (size_t idx=0; idx<text.size(); ++idx) {
		if (!iswprint(text[idx]))
			text[idx] = L'\uFFFD';
	}
	return text;
}

unsigned int utils::gentabs(const std::string& str) {
	int tabcount = 4 - (utils::strwidth(str) / 8);
	if (tabcount <= 0) {
		tabcount = 1;
	}
	return tabcount;
}

/* Like mkdir(), but creates ancestors (parent directories) if they don't
 * exist. */
int utils::mkdir_parents(const std::string& p, mode_t mode) {
	int result = -1;

	/* Have to copy the path because we're going to modify it */
	const size_t length = p.length() + 1;
	char* pathname = (char*)malloc(length);
	strncpy(pathname, p.c_str(), length);
	/* This pointer will run through the whole string looking for '/'.
	 * We move it by one if path starts with slash because if we don't, the
	 * first call to access() will fail (because of empty path) */
	char* curr = pathname + (*pathname == '/' ? 1 : 0);

	while (*curr) {
		if (*curr == '/') {
			*curr = '\0';
			result = access(pathname, F_OK);
			if (result == -1) {
				result = mkdir(pathname, mode);
				if (result != 0)
					break;
			}
			*curr = '/';
		}
		curr++;
	}

	if (result == 0) mkdir(p.c_str(), mode);

	free(pathname);
	return result;
}

std::string utils::make_title(const std::string& const_url) {
	/* Sometimes it is possible to construct the title from the URL
	 * This attempts to do just that. eg: http://domain.com/story/yy/mm/dd/title-with-dashes?a=b
	*/
	std::string url = (std::string&) const_url;
	// Strip out trailing slashes
	while (url.length() > 0  &&  url.back() == '/') {
		url.erase(url.length() - 1);
	}
	// get to the final part of the URI's path
	std::string::size_type pos_of_slash = url.find_last_of('/');
	// extract just the juicy part 'title-with-dashes?a=b'
	std::string path = url.substr(pos_of_slash+1);
	// find where query part of URI starts
	std::string::size_type pos_of_qmrk = path.find_first_of('?');
	//throw away the query part 'title-with-dashes'
	std::string title = path.substr(0,pos_of_qmrk);
	//Throw away common webpage suffixes: .html, .php, .aspx, .htm
	std::regex rx("\\.html$|\\.htm$|\\.php$|\\.aspx$");
	title = std::regex_replace(title,rx,"");
	// if there is nothing left, just give up
	if (title.empty())
		return title;
	// 'title with dashes'
	std::replace(title.begin(), title.end(), '-', ' ');
	std::replace(title.begin(), title.end(), '_', ' ');
	//'Title with dashes'
	if (title.at(0)>= 'a' && title.at(0)<= 'z') {
		title[0] -= 'a' - 'A';
	}
	// Un-escape any percent-encoding, e.g. "It%27s%202017%21" -> "It's 2017!"
	auto const result = xmlURIUnescapeString(title.c_str(), 0, nullptr);
	if (result) {
		title = result;
		xmlFree(result);
	}
	return title;
}

int utils::run_interactively(
		const std::string& command, const std::string& caller)
{
	LOG(level::DEBUG, "%s: running `%s'", caller, command);

	int status = ::system(command.c_str());

	if (status == -1) {
		LOG(level::DEBUG, "%s: couldn't create a child process", caller);
	} else if (status == 127) {
		LOG(level::DEBUG, "%s: couldn't run shell", caller);
	}

	return status;
}

std::string utils::getcwd() {
	char cwdtmp[MAXPATHLEN];

	if (::getcwd(cwdtmp, sizeof(cwdtmp)) == nullptr) {
		strncpy(cwdtmp, strerror(errno), MAXPATHLEN);
	}

	return std::string(cwdtmp);
}

void utils::remove_soft_hyphens(std::string& text) {
	/* Remove all soft-hyphens as they can behave unpredictably (see
	 * https://github.com/akrennmair/newsbeuter/issues/259#issuecomment-259609490)
	 * and inadvertently render as hyphens */

	std::string::size_type pos = text.find("\u00AD");
	while (pos != std::string::npos) {
		text.erase(pos, 2);
		pos = text.find("\u00AD", pos);
	}
}


/*
 * See http://curl.haxx.se/libcurl/c/libcurl-tutorial.html#Multi-threading for a reason why we do this.
 *
 * These callbacks are deprecated as of OpenSSL 1.1.0; see the changelog:
 * https://www.openssl.org/news/changelog.html#x6
 */

#if HAVE_OPENSSL && OPENSSL_VERSION_NUMBER < 0x01010000fL
static std::mutex * openssl_mutexes = nullptr;
static int openssl_mutexes_size = 0;

static void openssl_mth_locking_function(int mode, int n, const char * file, int line) {
	if (n < 0 || n >= openssl_mutexes_size) {
		LOG(level::ERROR,"openssl_mth_locking_function: index is out of bounds (called by %s:%d)", file, line);
		return;
	}
	if (mode & CRYPTO_LOCK) {
		LOG(level::DEBUG, "OpenSSL lock %d: %s:%d", n, file, line);
		openssl_mutexes[n].lock();
	} else {
		LOG(level::DEBUG, "OpenSSL unlock %d: %s:%d", n, file, line);
		openssl_mutexes[n].unlock();
	}
}

static unsigned long openssl_mth_id_function(void) {
	return (unsigned long)pthread_self();
}
#endif

void utils::initialize_ssl_implementation(void) {
#if HAVE_OPENSSL && OPENSSL_VERSION_NUMBER < 0x01010000fL
	openssl_mutexes_size = CRYPTO_num_locks();
	openssl_mutexes = new std::mutex[openssl_mutexes_size];
	CRYPTO_set_id_callback(openssl_mth_id_function);
	CRYPTO_set_locking_callback(openssl_mth_locking_function);
#endif

#if HAVE_GCRYPT
	gcry_control (GCRYCTL_SET_THREAD_CBS, &gcry_threads_pthread);
	gnutls_global_init();
#endif
}

}