src/xz/message.c

///////////////////////////////////////////////////////////////////////////////
//
/// \file       message.c
/// \brief      Printing messages
//
//  Author:     Lasse Collin
//
//  This file has been put into the public domain.
//  You can do whatever you want with this file.
//
///////////////////////////////////////////////////////////////////////////////

#include "private.h"

#include <stdarg.h>


/// Number of the current file
static unsigned int files_pos = 0;

/// Total number of input files; zero if unknown.
static unsigned int files_total;

/// Verbosity level
static enum message_verbosity verbosity = V_WARNING;

/// Filename which we will print with the verbose messages
static const char *filename;

/// True once the a filename has been printed to stderr as part of progress
/// message. If automatic progress updating isn't enabled, this becomes true
/// after the first progress message has been printed due to user sending
/// SIGINFO, SIGUSR1, or SIGALRM. Once this variable is true, we will print
/// an empty line before the next filename to make the output more readable.
static bool first_filename_printed = false;

/// This is set to true when we have printed the current filename to stderr
/// as part of a progress message. This variable is useful only if not
/// updating progress automatically: if user sends many SIGINFO, SIGUSR1, or
/// SIGALRM signals, we won't print the name of the same file multiple times.
static bool current_filename_printed = false;

/// True if we should print progress indicator and update it automatically
/// if also verbose >= V_VERBOSE.
static bool progress_automatic;

/// True if message_progress_start() has been called but
/// message_progress_end() hasn't been called yet.
static bool progress_started = false;

/// This is true when a progress message was printed and the cursor is still
/// on the same line with the progress message. In that case, a newline has
/// to be printed before any error messages.
static bool progress_active = false;

/// Pointer to lzma_stream used to do the encoding or decoding.
static lzma_stream *progress_strm;

/// This is true if we are in passthru mode (not actually compressing or
/// decompressing) and thus cannot use lzma_get_progress(progress_strm, ...).
/// That is, we are using coder_passthru() in coder.c.
static bool progress_is_from_passthru;

/// Expected size of the input stream is needed to show completion percentage
/// and estimate remaining time.
static uint64_t expected_in_size;


// Use alarm() and SIGALRM when they are supported. This has two minor
// advantages over the alternative of polling gettimeofday():
//  - It is possible for the user to send SIGINFO, SIGUSR1, or SIGALRM to
//    get intermediate progress information even when --verbose wasn't used
//    or stderr is not a terminal.
//  - alarm() + SIGALRM seems to have slightly less overhead than polling
//    gettimeofday().
#ifdef SIGALRM

const int message_progress_sigs[] = {
	SIGALRM,
#ifdef SIGINFO
	SIGINFO,
#endif
#ifdef SIGUSR1
	SIGUSR1,
#endif
	0
};

/// The signal handler for SIGALRM sets this to true. It is set back to false
/// once the progress message has been updated.
static volatile sig_atomic_t progress_needs_updating = false;

/// Signal handler for SIGALRM
static void
progress_signal_handler(int sig lzma_attribute((__unused__)))
{
	progress_needs_updating = true;
	return;
}

#else

/// This is true when progress message printing is wanted. Using the same
/// variable name as above to avoid some ifdefs.
static bool progress_needs_updating = false;

/// Elapsed time when the next progress message update should be done.
static uint64_t progress_next_update;

#endif


extern void
message_init(void)
{
	// If --verbose is used, we use a progress indicator if and only
	// if stderr is a terminal. If stderr is not a terminal, we print
	// verbose information only after finishing the file. As a special
	// exception, even if --verbose was not used, user can send SIGALRM
	// to make us print progress information once without automatic
	// updating.
	progress_automatic = isatty(STDERR_FILENO);

	// Commented out because COLUMNS is rarely exported to environment.
	// Most users have at least 80 columns anyway, let's think something
	// fancy here if enough people complain.
/*
	if (progress_automatic) {
		// stderr is a terminal. Check the COLUMNS environment
		// variable to see if the terminal is wide enough. If COLUMNS
		// doesn't exist or it has some unparsable value, we assume
		// that the terminal is wide enough.
		const char *columns_str = getenv("COLUMNS");
		if (columns_str != NULL) {
			char *endptr;
			const long columns = strtol(columns_str, &endptr, 10);
			if (*endptr != '\0' || columns < 80)
				progress_automatic = false;
		}
	}
*/

#ifdef SIGALRM
	// Establish the signal handlers which set a flag to tell us that
	// progress info should be updated.
	struct sigaction sa;
	sigemptyset(&sa.sa_mask);
	sa.sa_flags = 0;
	sa.sa_handler = &progress_signal_handler;

	for (size_t i = 0; message_progress_sigs[i] != 0; ++i)
		if (sigaction(message_progress_sigs[i], &sa, NULL))
			message_signal_handler();
#endif

	return;
}


extern void
message_verbosity_increase(void)
{
	if (verbosity < V_DEBUG)
		++verbosity;

	return;
}


extern void
message_verbosity_decrease(void)
{
	if (verbosity > V_SILENT)
		--verbosity;

	return;
}


extern enum message_verbosity
message_verbosity_get(void)
{
	return verbosity;
}


extern void
message_set_files(unsigned int files)
{
	files_total = files;
	return;
}


/// Prints the name of the current file if it hasn't been printed already,
/// except if we are processing exactly one stream from stdin to stdout.
/// I think it looks nicer to not print "(stdin)" when --verbose is used
/// in a pipe and no other files are processed.
static void
print_filename(void)
{
	if (!opt_robot && (files_total != 1 || filename != stdin_filename)) {
		signals_block();

		FILE *file = opt_mode == MODE_LIST ? stdout : stderr;

		// If a file was already processed, put an empty line
		// before the next filename to improve readability.
		if (first_filename_printed)
			fputc('\n', file);

		first_filename_printed = true;
		current_filename_printed = true;

		// If we don't know how many files there will be due
		// to usage of --files or --files0.
		if (files_total == 0)
			fprintf(file, "%s (%u)\n", filename,
					files_pos);
		else
			fprintf(file, "%s (%u/%u)\n", filename,
					files_pos, files_total);

		signals_unblock();
	}

	return;
}


extern void
message_filename(const char *src_name)
{
	// Start numbering the files starting from one.
	++files_pos;
	filename = src_name;

	if (verbosity >= V_VERBOSE
			&& (progress_automatic || opt_mode == MODE_LIST))
		print_filename();
	else
		current_filename_printed = false;

	return;
}


extern void
message_progress_start(lzma_stream *strm, bool is_passthru, uint64_t in_size)
{
	// Store the pointer to the lzma_stream used to do the coding.
	// It is needed to find out the position in the stream.
	progress_strm = strm;
	progress_is_from_passthru = is_passthru;

	// Store the expected size of the file. If we aren't printing any
	// statistics, then is will be unused. But since it is possible
	// that the user sends us a signal to show statistics, we need
	// to have it available anyway.
	expected_in_size = in_size;

	// Indicate that progress info may need to be printed before
	// printing error messages.
	progress_started = true;

	// If progress indicator is wanted, print the filename and possibly
	// the file count now.
	if (verbosity >= V_VERBOSE && progress_automatic) {
		// Start the timer to display the first progress message
		// after one second. An alternative would be to show the
		// first message almost immediately, but delaying by one
		// second looks better to me, since extremely early
		// progress info is pretty much useless.
#ifdef SIGALRM
		// First disable a possibly existing alarm.
		alarm(0);
		progress_needs_updating = false;
		alarm(1);
#else
		progress_needs_updating = true;
		progress_next_update = 1000;
#endif
	}

	return;
}


/// Make the string indicating completion percentage.
static const char *
progress_percentage(uint64_t in_pos)
{
	// If the size of the input file is unknown or the size told us is
	// clearly wrong since we have processed more data than the alleged
	// size of the file, show a static string indicating that we have
	// no idea of the completion percentage.
	if (expected_in_size == 0 || in_pos > expected_in_size)
		return "--- %";

	// Never show 100.0 % before we actually are finished.
	double percentage = (double)(in_pos) / (double)(expected_in_size)
			* 99.9;

	// Use big enough buffer to hold e.g. a multibyte decimal point.
	static char buf[16];
	snprintf(buf, sizeof(buf), "%.1f %%", percentage);

	return buf;
}


/// Make the string containing the amount of input processed, amount of
/// output produced, and the compression ratio.
static const char *
progress_sizes(uint64_t compressed_pos, uint64_t uncompressed_pos, bool final)
{
	// Use big enough buffer to hold e.g. a multibyte thousand separators.
	static char buf[128];
	char *pos = buf;
	size_t left = sizeof(buf);

	// Print the sizes. If this the final message, use more reasonable
	// units than MiB if the file was small.
	const enum nicestr_unit unit_min = final ? NICESTR_B : NICESTR_MIB;
	my_snprintf(&pos, &left, "%s / %s",
			uint64_to_nicestr(compressed_pos,
				unit_min, NICESTR_TIB, false, 0),
			uint64_to_nicestr(uncompressed_pos,
				unit_min, NICESTR_TIB, false, 1));

	// Avoid division by zero. If we cannot calculate the ratio, set
	// it to some nice number greater than 10.0 so that it gets caught
	// in the next if-clause.
	const double ratio = uncompressed_pos > 0
			? (double)(compressed_pos) / (double)(uncompressed_pos)
			: 16.0;

	// If the ratio is very bad, just indicate that it is greater than
	// 9.999. This way the length of the ratio field stays fixed.
	if (ratio > 9.999)
		snprintf(pos, left, " > %.3f", 9.999);
	else
		snprintf(pos, left, " = %.3f", ratio);

	return buf;
}


/// Make the string containing the processing speed of uncompressed data.
static const char *
progress_speed(uint64_t uncompressed_pos, uint64_t elapsed)
{
	// Don't print the speed immediately, since the early values look
	// somewhat random.
	if (elapsed < 3000)
		return "";

	static const char unit[][8] = {
		"KiB/s",
		"MiB/s",
		"GiB/s",
	};

	size_t unit_index = 0;

	// Calculate the speed as KiB/s.
	double speed = (double)(uncompressed_pos)
			/ ((double)(elapsed) * (1024.0 / 1000.0));

	// Adjust the unit of the speed if needed.
	while (speed > 999.0) {
		speed /= 1024.0;
		if (++unit_index == ARRAY_SIZE(unit))
			return ""; // Way too fast ;-)
	}

	// Use decimal point only if the number is small. Examples:
	//  - 0.1 KiB/s
	//  - 9.9 KiB/s
	//  - 99 KiB/s
	//  - 999 KiB/s
	// Use big enough buffer to hold e.g. a multibyte decimal point.
	static char buf[16];
	snprintf(buf, sizeof(buf), "%.*f %s",
			speed > 9.9 ? 0 : 1, speed, unit[unit_index]);
	return buf;
}


/// Make a string indicating elapsed time. The format is either
/// M:SS or H:MM:SS depending on if the time is an hour or more.
static const char *
progress_time(uint64_t mseconds)
{
	// 9999 hours = 416 days
	static char buf[sizeof("9999:59:59")];

	// 32-bit variable is enough for elapsed time (136 years).
	uint32_t seconds = (uint32_t)(mseconds / 1000);

	// Don't show anything if the time is zero or ridiculously big.
	if (seconds == 0 || seconds > ((9999 * 60) + 59) * 60 + 59)
		return "";

	uint32_t minutes = seconds / 60;
	seconds %= 60;

	if (minutes >= 60) {
		const uint32_t hours = minutes / 60;
		minutes %= 60;
		snprintf(buf, sizeof(buf),
				"%" PRIu32 ":%02" PRIu32 ":%02" PRIu32,
				hours, minutes, seconds);
	} else {
		snprintf(buf, sizeof(buf), "%" PRIu32 ":%02" PRIu32,
				minutes, seconds);
	}

	return buf;
}


/// Return a string containing estimated remaining time when
/// reasonably possible.
static const char *
progress_remaining(uint64_t in_pos, uint64_t elapsed)
{
	// Don't show the estimated remaining time when it wouldn't
	// make sense:
	//  - Input size is unknown.
	//  - Input has grown bigger since we started (de)compressing.
	//  - We haven't processed much data yet, so estimate would be
	//    too inaccurate.
	//  - Only a few seconds has passed since we started (de)compressing,
	//    so estimate would be too inaccurate.
	if (expected_in_size == 0 || in_pos > expected_in_size
			|| in_pos < (UINT64_C(1) << 19) || elapsed < 8000)
		return "";

	// Calculate the estimate. Don't give an estimate of zero seconds,
	// since it is possible that all the input has been already passed
	// to the library, but there is still quite a bit of output pending.
	uint32_t remaining = (uint32_t)((double)(expected_in_size - in_pos)
			* ((double)(elapsed) / 1000.0) / (double)(in_pos));
	if (remaining < 1)
		remaining = 1;

	static char buf[sizeof("9 h 55 min")];

	// Select appropriate precision for the estimated remaining time.
	if (remaining <= 10) {
		// A maximum of 10 seconds remaining.
		// Show the number of seconds as is.
		snprintf(buf, sizeof(buf), "%" PRIu32 " s", remaining);

	} else if (remaining <= 50) {
		// A maximum of 50 seconds remaining.
		// Round up to the next multiple of five seconds.
		remaining = (remaining + 4) / 5 * 5;
		snprintf(buf, sizeof(buf), "%" PRIu32 " s", remaining);

	} else if (remaining <= 590) {
		// A maximum of 9 minutes and 50 seconds remaining.
		// Round up to the next multiple of ten seconds.
		remaining = (remaining + 9) / 10 * 10;
		snprintf(buf, sizeof(buf), "%" PRIu32 " min %" PRIu32 " s",
				remaining / 60, remaining % 60);

	} else if (remaining <= 59 * 60) {
		// A maximum of 59 minutes remaining.
		// Round up to the next multiple of a minute.
		remaining = (remaining + 59) / 60;
		snprintf(buf, sizeof(buf), "%" PRIu32 " min", remaining);

	} else if (remaining <= 9 * 3600 + 50 * 60) {
		// A maximum of 9 hours and 50 minutes left.
		// Round up to the next multiple of ten minutes.
		remaining = (remaining + 599) / 600 * 10;
		snprintf(buf, sizeof(buf), "%" PRIu32 " h %" PRIu32 " min",
				remaining / 60, remaining % 60);

	} else if (remaining <= 23 * 3600) {
		// A maximum of 23 hours remaining.
		// Round up to the next multiple of an hour.
		remaining = (remaining + 3599) / 3600;
		snprintf(buf, sizeof(buf), "%" PRIu32 " h", remaining);

	} else if (remaining <= 9 * 24 * 3600 + 23 * 3600) {
		// A maximum of 9 days and 23 hours remaining.
		// Round up to the next multiple of an hour.
		remaining = (remaining + 3599) / 3600;
		snprintf(buf, sizeof(buf), "%" PRIu32 " d %" PRIu32 " h",
				remaining / 24, remaining % 24);

	} else if (remaining <= 999 * 24 * 3600) {
		// A maximum of 999 days remaining. ;-)
		// Round up to the next multiple of a day.
		remaining = (remaining + 24 * 3600 - 1) / (24 * 3600);
		snprintf(buf, sizeof(buf), "%" PRIu32 " d", remaining);

	} else {
		// The estimated remaining time is too big. Don't show it.
		return "";
	}

	return buf;
}


/// Get how much uncompressed and compressed data has been processed.
static void
progress_pos(uint64_t *in_pos,
		uint64_t *compressed_pos, uint64_t *uncompressed_pos)
{
	uint64_t out_pos;
	if (progress_is_from_passthru) {
		// In passthru mode the progress info is in total_in/out but
		// the *progress_strm itself isn't initialized and thus we
		// cannot use lzma_get_progress().
		*in_pos = progress_strm->total_in;
		out_pos = progress_strm->total_out;
	} else {
		lzma_get_progress(progress_strm, in_pos, &out_pos);
	}

	// It cannot have processed more input than it has been given.
	assert(*in_pos <= progress_strm->total_in);

	// It cannot have produced more output than it claims to have ready.
	assert(out_pos >= progress_strm->total_out);

	if (opt_mode == MODE_COMPRESS) {
		*compressed_pos = out_pos;
		*uncompressed_pos = *in_pos;
	} else {
		*compressed_pos = *in_pos;
		*uncompressed_pos = out_pos;
	}

	return;
}


extern void
message_progress_update(void)
{
	if (!progress_needs_updating)
		return;

	// Calculate how long we have been processing this file.
	const uint64_t elapsed = mytime_get_elapsed();

#ifndef SIGALRM
	if (progress_next_update > elapsed)
		return;

	progress_next_update = elapsed + 1000;
#endif

	// Get our current position in the stream.
	uint64_t in_pos;
	uint64_t compressed_pos;
	uint64_t uncompressed_pos;
	progress_pos(&in_pos, &compressed_pos, &uncompressed_pos);

	// Block signals so that fprintf() doesn't get interrupted.
	signals_block();

	// Print the filename if it hasn't been printed yet.
	if (!current_filename_printed)
		print_filename();

	// Print the actual progress message. The idea is that there is at
	// least three spaces between the fields in typical situations, but
	// even in rare situations there is at least one space.
	const char *cols[5] = {
		progress_percentage(in_pos),
		progress_sizes(compressed_pos, uncompressed_pos, false),
		progress_speed(uncompressed_pos, elapsed),
		progress_time(elapsed),
		progress_remaining(in_pos, elapsed),
	};
	fprintf(stderr, "\r %*s %*s   %*s %10s   %10s\r",
			tuklib_mbstr_fw(cols[0], 6), cols[0],
			tuklib_mbstr_fw(cols[1], 35), cols[1],
			tuklib_mbstr_fw(cols[2], 9), cols[2],
			cols[3],
			cols[4]);

#ifdef SIGALRM
	// Updating the progress info was finished. Reset
	// progress_needs_updating to wait for the next SIGALRM.
	//
	// NOTE: This has to be done before alarm(1) or with (very) bad
	// luck we could be setting this to false after the alarm has already
	// been triggered.
	progress_needs_updating = false;

	if (verbosity >= V_VERBOSE && progress_automatic) {
		// Mark that the progress indicator is active, so if an error
		// occurs, the error message gets printed cleanly.
		progress_active = true;

		// Restart the timer so that progress_needs_updating gets
		// set to true after about one second.
		alarm(1);
	} else {
		// The progress message was printed because user had sent us
		// SIGALRM. In this case, each progress message is printed
		// on its own line.
		fputc('\n', stderr);
	}
#else
	// When SIGALRM isn't supported and we get here, it's always due to
	// automatic progress update. We set progress_active here too like
	// described above.
	assert(verbosity >= V_VERBOSE);
	assert(progress_automatic);
	progress_active = true;
#endif

	signals_unblock();

	return;
}


static void
progress_flush(bool finished)
{
	if (!progress_started || verbosity < V_VERBOSE)
		return;

	uint64_t in_pos;
	uint64_t compressed_pos;
	uint64_t uncompressed_pos;
	progress_pos(&in_pos, &compressed_pos, &uncompressed_pos);

	// Avoid printing intermediate progress info if some error occurs
	// in the beginning of the stream. (If something goes wrong later in
	// the stream, it is sometimes useful to tell the user where the
	// error approximately occurred, especially if the error occurs
	// after a time-consuming operation.)
	if (!finished && !progress_active
			&& (compressed_pos == 0 || uncompressed_pos == 0))
		return;

	progress_active = false;

	const uint64_t elapsed = mytime_get_elapsed();

	signals_block();

	// When using the auto-updating progress indicator, the final
	// statistics are printed in the same format as the progress
	// indicator itself.
	if (progress_automatic) {
		const char *cols[5] = {
			finished ? "100 %" : progress_percentage(in_pos),
			progress_sizes(compressed_pos, uncompressed_pos, true),
			progress_speed(uncompressed_pos, elapsed),
			progress_time(elapsed),
			finished ? "" : progress_remaining(in_pos, elapsed),
		};
		fprintf(stderr, "\r %*s %*s   %*s %10s   %10s\n",
				tuklib_mbstr_fw(cols[0], 6), cols[0],
				tuklib_mbstr_fw(cols[1], 35), cols[1],
				tuklib_mbstr_fw(cols[2], 9), cols[2],
				cols[3],
				cols[4]);
	} else {
		// The filename is always printed.
		fprintf(stderr, "%s: ", filename);

		// Percentage is printed only if we didn't finish yet.
		if (!finished) {
			// Don't print the percentage when it isn't known
			// (starts with a dash).
			const char *percentage = progress_percentage(in_pos);
			if (percentage[0] != '-')
				fprintf(stderr, "%s, ", percentage);
		}

		// Size information is always printed.
		fprintf(stderr, "%s", progress_sizes(
				compressed_pos, uncompressed_pos, true));

		// The speed and elapsed time aren't always shown.
		const char *speed = progress_speed(uncompressed_pos, elapsed);
		if (speed[0] != '\0')
			fprintf(stderr, ", %s", speed);

		const char *elapsed_str = progress_time(elapsed);
		if (elapsed_str[0] != '\0')
			fprintf(stderr, ", %s", elapsed_str);

		fputc('\n', stderr);
	}

	signals_unblock();

	return;
}


extern void
message_progress_end(bool success)
{
	assert(progress_started);
	progress_flush(success);
	progress_started = false;
	return;
}


static void
vmessage(enum message_verbosity v, const char *fmt, va_list ap)
{
	if (v <= verbosity) {
		signals_block();

		progress_flush(false);

		// TRANSLATORS: This is the program name in the beginning
		// of the line in messages. Usually it becomes "xz: ".
		// This is a translatable string because French needs
		// a space before a colon.
		fprintf(stderr, _("%s: "), progname);
		vfprintf(stderr, fmt, ap);
		fputc('\n', stderr);

		signals_unblock();
	}

	return;
}


extern void
message(enum message_verbosity v, const char *fmt, ...)
{
	va_list ap;
	va_start(ap, fmt);
	vmessage(v, fmt, ap);
	va_end(ap);
	return;
}


extern void
message_warning(const char *fmt, ...)
{
	va_list ap;
	va_start(ap, fmt);
	vmessage(V_WARNING, fmt, ap);
	va_end(ap);

	set_exit_status(E_WARNING);
	return;
}


extern void
message_error(const char *fmt, ...)
{
	va_list ap;
	va_start(ap, fmt);
	vmessage(V_ERROR, fmt, ap);
	va_end(ap);

	set_exit_status(E_ERROR);
	return;
}


extern void
message_fatal(const char *fmt, ...)
{
	va_list ap;
	va_start(ap, fmt);
	vmessage(V_ERROR, fmt, ap);
	va_end(ap);

	tuklib_exit(E_ERROR, E_ERROR, false);
}


extern void
message_bug(void)
{
	message_fatal(_("Internal error (bug)"));
}


extern void
message_signal_handler(void)
{
	message_fatal(_("Cannot establish signal handlers"));
}


extern const char *
message_strm(lzma_ret code)
{
	switch (code) {
	case LZMA_NO_CHECK:
		return _("No integrity check; not verifying file integrity");

	case LZMA_UNSUPPORTED_CHECK:
		return _("Unsupported type of integrity check; "
				"not verifying file integrity");

	case LZMA_MEM_ERROR:
		return strerror(ENOMEM);

	case LZMA_MEMLIMIT_ERROR:
		return _("Memory usage limit reached");

	case LZMA_FORMAT_ERROR:
		return _("File format not recognized");

	case LZMA_OPTIONS_ERROR:
		return _("Unsupported options");

	case LZMA_DATA_ERROR:
		return _("Compressed data is corrupt");

	case LZMA_BUF_ERROR:
		return _("Unexpected end of input");

	case LZMA_OK:
	case LZMA_STREAM_END:
	case LZMA_GET_CHECK:
	case LZMA_PROG_ERROR:
		// Without "default", compiler will warn if new constants
		// are added to lzma_ret, it is not too easy to forget to
		// add the new constants to this function.
		break;
	}

	return _("Internal error (bug)");
}


extern void
message_mem_needed(enum message_verbosity v, uint64_t memusage)
{
	if (v > verbosity)
		return;

	// Convert memusage to MiB, rounding up to the next full MiB.
	// This way the user can always use the displayed usage as
	// the new memory usage limit. (If we rounded to the nearest,
	// the user might need to +1 MiB to get high enough limit.)
	memusage = round_up_to_mib(memusage);

	uint64_t memlimit = hardware_memlimit_get(opt_mode);

	// Handle the case when there is no memory usage limit.
	// This way we don't print a weird message with a huge number.
	if (memlimit == UINT64_MAX) {
		message(v, _("%s MiB of memory is required. "
				"The limiter is disabled."),
				uint64_to_str(memusage, 0));
		return;
	}

	// With US-ASCII:
	// 2^64 with thousand separators + " MiB" suffix + '\0' = 26 + 4 + 1
	// But there may be multibyte chars so reserve enough space.
	char memlimitstr[128];

	// Show the memory usage limit as MiB unless it is less than 1 MiB.
	// This way it's easy to notice errors where one has typed
	// --memory=123 instead of --memory=123MiB.
	if (memlimit < (UINT32_C(1) << 20)) {
		snprintf(memlimitstr, sizeof(memlimitstr), "%s B",
				uint64_to_str(memlimit, 1));
	} else {
		// Round up just like with memusage. If this function is
		// called for informational purposes (to just show the
		// current usage and limit), we should never show that
		// the usage is higher than the limit, which would give
		// a false impression that the memory usage limit isn't
		// properly enforced.
		snprintf(memlimitstr, sizeof(memlimitstr), "%s MiB",
				uint64_to_str(round_up_to_mib(memlimit), 1));
	}

	message(v, _("%s MiB of memory is required. The limit is %s."),
			uint64_to_str(memusage, 0), memlimitstr);

	return;
}


/// \brief      Convert uint32_t to a nice string for --lzma[12]=dict=SIZE
///
/// The idea is to use KiB or MiB suffix when possible.
static const char *
uint32_to_optstr(uint32_t num)
{
	static char buf[16];

	if ((num & ((UINT32_C(1) << 20) - 1)) == 0)
		snprintf(buf, sizeof(buf), "%" PRIu32 "MiB", num >> 20);
	else if ((num & ((UINT32_C(1) << 10) - 1)) == 0)
		snprintf(buf, sizeof(buf), "%" PRIu32 "KiB", num >> 10);
	else
		snprintf(buf, sizeof(buf), "%" PRIu32, num);

	return buf;
}


extern void
message_filters_to_str(char buf[FILTERS_STR_SIZE],
		const lzma_filter *filters, bool all_known)
{
	char *pos = buf;
	size_t left = FILTERS_STR_SIZE;

	for (size_t i = 0; filters[i].id != LZMA_VLI_UNKNOWN; ++i) {
		// Add the dashes for the filter option. A space is
		// needed after the first and later filters.
		my_snprintf(&pos, &left, "%s", i == 0 ? "--" : " --");

		switch (filters[i].id) {
		case LZMA_FILTER_LZMA1:
		case LZMA_FILTER_LZMA2: {
			const lzma_options_lzma *opt = filters[i].options;
			const char *mode = NULL;
			const char *mf = NULL;

			if (all_known) {
				switch (opt->mode) {
				case LZMA_MODE_FAST:
					mode = "fast";
					break;

				case LZMA_MODE_NORMAL:
					mode = "normal";
					break;

				default:
					mode = "UNKNOWN";
					break;
				}

				switch (opt->mf) {
				case LZMA_MF_HC3:
					mf = "hc3";
					break;

				case LZMA_MF_HC4:
					mf = "hc4";
					break;

				case LZMA_MF_BT2:
					mf = "bt2";
					break;

				case LZMA_MF_BT3:
					mf = "bt3";
					break;

				case LZMA_MF_BT4:
					mf = "bt4";
					break;

				default:
					mf = "UNKNOWN";
					break;
				}
			}

			// Add the filter name and dictionary size, which
			// is always known.
			my_snprintf(&pos, &left, "lzma%c=dict=%s",
					filters[i].id == LZMA_FILTER_LZMA2
						? '2' : '1',
					uint32_to_optstr(opt->dict_size));

			// With LZMA1 also lc/lp/pb are known when
			// decompressing, but this function is never
			// used to print information about .lzma headers.
			assert(filters[i].id == LZMA_FILTER_LZMA2
					|| all_known);

			// Print the rest of the options, which are known
			// only when compressing.
			if (all_known)
				my_snprintf(&pos, &left,
					",lc=%" PRIu32 ",lp=%" PRIu32
					",pb=%" PRIu32
					",mode=%s,nice=%" PRIu32 ",mf=%s"
					",depth=%" PRIu32,
					opt->lc, opt->lp, opt->pb,
					mode, opt->nice_len, mf, opt->depth);
			break;
		}

		case LZMA_FILTER_X86:
		case LZMA_FILTER_POWERPC:
		case LZMA_FILTER_IA64:
		case LZMA_FILTER_ARM:
		case LZMA_FILTER_ARMTHUMB:
		case LZMA_FILTER_SPARC: {
			static const char bcj_names[][9] = {
				"x86",
				"powerpc",
				"ia64",
				"arm",
				"armthumb",
				"sparc",
			};

			const lzma_options_bcj *opt = filters[i].options;
			my_snprintf(&pos, &left, "%s", bcj_names[filters[i].id
					- LZMA_FILTER_X86]);

			// Show the start offset only when really needed.
			if (opt != NULL && opt->start_offset != 0)
				my_snprintf(&pos, &left, "=start=%" PRIu32,
						opt->start_offset);

			break;
		}

		case LZMA_FILTER_DELTA: {
			const lzma_options_delta *opt = filters[i].options;
			my_snprintf(&pos, &left, "delta=dist=%" PRIu32,
					opt->dist);
			break;
		}

		default:
			// This should be possible only if liblzma is
			// newer than the xz tool.
			my_snprintf(&pos, &left, "UNKNOWN");
			break;
		}
	}

	return;
}


extern void
message_filters_show(enum message_verbosity v, const lzma_filter *filters)
{
	if (v > verbosity)
		return;

	char buf[FILTERS_STR_SIZE];
	message_filters_to_str(buf, filters, true);
	fprintf(stderr, _("%s: Filter chain: %s\n"), progname, buf);
	return;
}


extern void
message_try_help(void)
{
	// Print this with V_WARNING instead of V_ERROR to prevent it from
	// showing up when --quiet has been specified.
	message(V_WARNING, _("Try `%s --help' for more information."),
			progname);
	return;
}


extern void
message_version(void)
{
	// It is possible that liblzma version is different than the command
	// line tool version, so print both.
	if (opt_robot) {
		printf("XZ_VERSION=%" PRIu32 "\nLIBLZMA_VERSION=%" PRIu32 "\n",
				LZMA_VERSION, lzma_version_number());
	} else {
		printf("xz (" PACKAGE_NAME ") " LZMA_VERSION_STRING "\n");
		printf("liblzma %s\n", lzma_version_string());
	}

	tuklib_exit(E_SUCCESS, E_ERROR, verbosity != V_SILENT);
}


extern void
message_help(bool long_help)
{
	printf(_("Usage: %s [OPTION]... [FILE]...\n"
			"Compress or decompress FILEs in the .xz format.\n\n"),
			progname);

	// NOTE: The short help doesn't currently have options that
	// take arguments.
	if (long_help)
		puts(_("Mandatory arguments to long options are mandatory "
				"for short options too.\n"));

	if (long_help)
		puts(_(" Operation mode:\n"));

	puts(_(
"  -z, --compress      force compression\n"
"  -d, --decompress    force decompression\n"
"  -t, --test          test compressed file integrity\n"
"  -l, --list          list information about .xz files"));

	if (long_help)
		puts(_("\n Operation modifiers:\n"));

	puts(_(
"  -k, --keep          keep (don't delete) input files\n"
"  -f, --force         force overwrite of output file and (de)compress links\n"
"  -c, --stdout        write to standard output and don't delete input files"));

	if (long_help) {
		puts(_(
"      --single-stream decompress only the first stream, and silently\n"
"                      ignore possible remaining input data"));
		puts(_(
"      --no-sparse     do not create sparse files when decompressing\n"
"  -S, --suffix=.SUF   use the suffix `.SUF' on compressed files\n"
"      --files[=FILE]  read filenames to process from FILE; if FILE is\n"
"                      omitted, filenames are read from the standard input;\n"
"                      filenames must be terminated with the newline character\n"
"      --files0[=FILE] like --files but use the null character as terminator"));
	}

	if (long_help) {
		puts(_("\n Basic file format and compression options:\n"));
		puts(_(
"  -F, --format=FMT    file format to encode or decode; possible values are\n"
"                      `auto' (default), `xz', `lzma', and `raw'\n"
"  -C, --check=CHECK   integrity check type: `none' (use with caution),\n"
"                      `crc32', `crc64' (default), or `sha256'"));
		puts(_(
"      --ignore-check  don't verify the integrity check when decompressing"));
	}

	puts(_(
"  -0 ... -9           compression preset; default is 6; take compressor *and*\n"
"                      decompressor memory usage into account before using 7-9!"));

	puts(_(
"  -e, --extreme       try to improve compression ratio by using more CPU time;\n"
"                      does not affect decompressor memory requirements"));

	puts(_(
"  -T, --threads=NUM   use at most NUM threads; the default is 1; set to 0\n"
"                      to use as many threads as there are processor cores"));

	if (long_help) {
		puts(_(
"      --block-size=SIZE\n"
"                      start a new .xz block after every SIZE bytes of input;\n"
"                      use this to set the block size for threaded compression"));
		puts(_(
"      --block-list=SIZES\n"
"                      start a new .xz block after the given comma-separated\n"
"                      intervals of uncompressed data"));
		puts(_(
"      --flush-timeout=TIMEOUT\n"
"                      when compressing, if more than TIMEOUT milliseconds has\n"
"                      passed since the previous flush and reading more input\n"
"                      would block, all pending data is flushed out"
		));
		puts(_( // xgettext:no-c-format
"      --memlimit-compress=LIMIT\n"
"      --memlimit-decompress=LIMIT\n"
"  -M, --memlimit=LIMIT\n"
"                      set memory usage limit for compression, decompression,\n"
"                      or both; LIMIT is in bytes, % of RAM, or 0 for defaults"));

		puts(_(
"      --no-adjust     if compression settings exceed the memory usage limit,\n"
"                      give an error instead of adjusting the settings downwards"));
	}

	if (long_help) {
		puts(_(
"\n Custom filter chain for compression (alternative for using presets):"));

#if defined(HAVE_ENCODER_LZMA1) || defined(HAVE_DECODER_LZMA1) \
		|| defined(HAVE_ENCODER_LZMA2) || defined(HAVE_DECODER_LZMA2)
		// TRANSLATORS: The word "literal" in "literal context bits"
		// means how many "context bits" to use when encoding
		// literals. A literal is a single 8-bit byte. It doesn't
		// mean "literally" here.
		puts(_(
"\n"
"  --lzma1[=OPTS]      LZMA1 or LZMA2; OPTS is a comma-separated list of zero or\n"
"  --lzma2[=OPTS]      more of the following options (valid values; default):\n"
"                        preset=PRE reset options to a preset (0-9[e])\n"
"                        dict=NUM   dictionary size (4KiB - 1536MiB; 8MiB)\n"
"                        lc=NUM     number of literal context bits (0-4; 3)\n"
"                        lp=NUM     number of literal position bits (0-4; 0)\n"
"                        pb=NUM     number of position bits (0-4; 2)\n"
"                        mode=MODE  compression mode (fast, normal; normal)\n"
"                        nice=NUM   nice length of a match (2-273; 64)\n"
"                        mf=NAME    match finder (hc3, hc4, bt2, bt3, bt4; bt4)\n"
"                        depth=NUM  maximum search depth; 0=automatic (default)"));
#endif

		puts(_(
"\n"
"  --x86[=OPTS]        x86 BCJ filter (32-bit and 64-bit)\n"
"  --powerpc[=OPTS]    PowerPC BCJ filter (big endian only)\n"
"  --ia64[=OPTS]       IA-64 (Itanium) BCJ filter\n"
"  --arm[=OPTS]        ARM BCJ filter (little endian only)\n"
"  --armthumb[=OPTS]   ARM-Thumb BCJ filter (little endian only)\n"
"  --sparc[=OPTS]      SPARC BCJ filter\n"
"                      Valid OPTS for all BCJ filters:\n"
"                        start=NUM  start offset for conversions (default=0)"));

#if defined(HAVE_ENCODER_DELTA) || defined(HAVE_DECODER_DELTA)
		puts(_(
"\n"
"  --delta[=OPTS]      Delta filter; valid OPTS (valid values; default):\n"
"                        dist=NUM   distance between bytes being subtracted\n"
"                                   from each other (1-256; 1)"));
#endif
	}

	if (long_help)
		puts(_("\n Other options:\n"));

	puts(_(
"  -q, --quiet         suppress warnings; specify twice to suppress errors too\n"
"  -v, --verbose       be verbose; specify twice for even more verbose"));

	if (long_help) {
		puts(_(
"  -Q, --no-warn       make warnings not affect the exit status"));
		puts(_(
"      --robot         use machine-parsable messages (useful for scripts)"));
		puts("");
		puts(_(
"      --info-memory   display the total amount of RAM and the currently active\n"
"                      memory usage limits, and exit"));
		puts(_(
"  -h, --help          display the short help (lists only the basic options)\n"
"  -H, --long-help     display this long help and exit"));
	} else {
		puts(_(
"  -h, --help          display this short help and exit\n"
"  -H, --long-help     display the long help (lists also the advanced options)"));
	}

	puts(_(
"  -V, --version       display the version number and exit"));

	puts(_("\nWith no FILE, or when FILE is -, read standard input.\n"));

	// TRANSLATORS: This message indicates the bug reporting address
	// for this package. Please add _another line_ saying
	// "Report translation bugs to <...>\n" with the email or WWW
	// address for translation bugs. Thanks.
	printf(_("Report bugs to <%s> (in English or Finnish).\n"),
			PACKAGE_BUGREPORT);
	printf(_("%s home page: <%s>\n"), PACKAGE_NAME, PACKAGE_URL);

#if LZMA_VERSION_STABILITY != LZMA_VERSION_STABILITY_STABLE
	puts(_(
"THIS IS A DEVELOPMENT VERSION NOT INTENDED FOR PRODUCTION USE."));
#endif

	tuklib_exit(E_SUCCESS, E_ERROR, verbosity != V_SILENT);
}