// SPDX-License-Identifier: 0BSD
///////////////////////////////////////////////////////////////////////////////
//
/// \file lzma_encoder.c
/// \brief LZMA encoder
///
// Authors: Igor Pavlov
// Lasse Collin
//
///////////////////////////////////////////////////////////////////////////////
#include "lzma2_encoder.h"
#include "lzma_encoder_private.h"
#include "fastpos.h"
/////////////
// Literal //
/////////////
static inline void
literal_matched(lzma_range_encoder *rc, probability *subcoder,
uint32_t match_byte, uint32_t symbol)
{
uint32_t offset = 0x100;
symbol += UINT32_C(1) << 8;
do {
match_byte <<= 1;
const uint32_t match_bit = match_byte & offset;
const uint32_t subcoder_index
= offset + match_bit + (symbol >> 8);
const uint32_t bit = (symbol >> 7) & 1;
rc_bit(rc, &subcoder[subcoder_index], bit);
symbol <<= 1;
offset &= ~(match_byte ^ symbol);
} while (symbol < (UINT32_C(1) << 16));
}
static inline void
literal(lzma_lzma1_encoder *coder, lzma_mf *mf, uint32_t position)
{
// Locate the literal byte to be encoded and the subcoder.
const uint8_t cur_byte = mf->buffer[
mf->read_pos - mf->read_ahead];
probability *subcoder = literal_subcoder(coder->literal,
coder->literal_context_bits, coder->literal_pos_mask,
position, mf->buffer[mf->read_pos - mf->read_ahead - 1]);
if (is_literal_state(coder->state)) {
// Previous LZMA-symbol was a literal. Encode a normal
// literal without a match byte.
update_literal_normal(coder->state);
rc_bittree(&coder->rc, subcoder, 8, cur_byte);
} else {
// Previous LZMA-symbol was a match. Use the last byte of
// the match as a "match byte". That is, compare the bits
// of the current literal and the match byte.
update_literal_matched(coder->state);
const uint8_t match_byte = mf->buffer[
mf->read_pos - coder->reps[0] - 1
- mf->read_ahead];
literal_matched(&coder->rc, subcoder, match_byte, cur_byte);
}
}
//////////////////
// Match length //
//////////////////
static void
length_update_prices(lzma_length_encoder *lc, const uint32_t pos_state)
{
const uint32_t table_size = lc->table_size;
lc->counters[pos_state] = table_size;
const uint32_t a0 = rc_bit_0_price(lc->choice);
const uint32_t a1 = rc_bit_1_price(lc->choice);
const uint32_t b0 = a1 + rc_bit_0_price(lc->choice2);
const uint32_t b1 = a1 + rc_bit_1_price(lc->choice2);
uint32_t *const prices = lc->prices[pos_state];
uint32_t i;
for (i = 0; i < table_size && i < LEN_LOW_SYMBOLS; ++i)
prices[i] = a0 + rc_bittree_price(lc->low[pos_state],
LEN_LOW_BITS, i);
for (; i < table_size && i < LEN_LOW_SYMBOLS + LEN_MID_SYMBOLS; ++i)
prices[i] = b0 + rc_bittree_price(lc->mid[pos_state],
LEN_MID_BITS, i - LEN_LOW_SYMBOLS);
for (; i < table_size; ++i)
prices[i] = b1 + rc_bittree_price(lc->high, LEN_HIGH_BITS,
i - LEN_LOW_SYMBOLS - LEN_MID_SYMBOLS);
return;
}
static inline void
length(lzma_range_encoder *rc, lzma_length_encoder *lc,
const uint32_t pos_state, uint32_t len, const bool fast_mode)
{
assert(len <= MATCH_LEN_MAX);
len -= MATCH_LEN_MIN;
if (len < LEN_LOW_SYMBOLS) {
rc_bit(rc, &lc->choice, 0);
rc_bittree(rc, lc->low[pos_state], LEN_LOW_BITS, len);
} else {
rc_bit(rc, &lc->choice, 1);
len -= LEN_LOW_SYMBOLS;
if (len < LEN_MID_SYMBOLS) {
rc_bit(rc, &lc->choice2, 0);
rc_bittree(rc, lc->mid[pos_state], LEN_MID_BITS, len);
} else {
rc_bit(rc, &lc->choice2, 1);
len -= LEN_MID_SYMBOLS;
rc_bittree(rc, lc->high, LEN_HIGH_BITS, len);
}
}
// Only getoptimum uses the prices so don't update the table when
// in fast mode.
if (!fast_mode)
if (--lc->counters[pos_state] == 0)
length_update_prices(lc, pos_state);
}
///////////
// Match //
///////////
static inline void
match(lzma_lzma1_encoder *coder, const uint32_t pos_state,
const uint32_t distance, const uint32_t len)
{
update_match(coder->state);
length(&coder->rc, &coder->match_len_encoder, pos_state, len,
coder->fast_mode);
const uint32_t dist_slot = get_dist_slot(distance);
const uint32_t dist_state = get_dist_state(len);
rc_bittree(&coder->rc, coder->dist_slot[dist_state],
DIST_SLOT_BITS, dist_slot);
if (dist_slot >= DIST_MODEL_START) {
const uint32_t footer_bits = (dist_slot >> 1) - 1;
const uint32_t base = (2 | (dist_slot & 1)) << footer_bits;
const uint32_t dist_reduced = distance - base;
if (dist_slot < DIST_MODEL_END) {
// Careful here: base - dist_slot - 1 can be -1, but
// rc_bittree_reverse starts at probs[1], not probs[0].
rc_bittree_reverse(&coder->rc,
coder->dist_special + base - dist_slot - 1,
footer_bits, dist_reduced);
} else {
rc_direct(&coder->rc, dist_reduced >> ALIGN_BITS,
footer_bits - ALIGN_BITS);
rc_bittree_reverse(
&coder->rc, coder->dist_align,
ALIGN_BITS, dist_reduced & ALIGN_MASK);
++coder->align_price_count;
}
}
coder->reps[3] = coder->reps[2];
coder->reps[2] = coder->reps[1];
coder->reps[1] = coder->reps[0];
coder->reps[0] = distance;
++coder->match_price_count;
}
////////////////////
// Repeated match //
////////////////////
static inline void
rep_match(lzma_lzma1_encoder *coder, const uint32_t pos_state,
const uint32_t rep, const uint32_t len)
{
if (rep == 0) {
rc_bit(&coder->rc, &coder->is_rep0[coder->state], 0);
rc_bit(&coder->rc,
&coder->is_rep0_long[coder->state][pos_state],
len != 1);
} else {
const uint32_t distance = coder->reps[rep];
rc_bit(&coder->rc, &coder->is_rep0[coder->state], 1);
if (rep == 1) {
rc_bit(&coder->rc, &coder->is_rep1[coder->state], 0);
} else {
rc_bit(&coder->rc, &coder->is_rep1[coder->state], 1);
rc_bit(&coder->rc, &coder->is_rep2[coder->state],
rep - 2);
if (rep == 3)
coder->reps[3] = coder->reps[2];
coder->reps[2] = coder->reps[1];
}
coder->reps[1] = coder->reps[0];
coder->reps[0] = distance;
}
if (len == 1) {
update_short_rep(coder->state);
} else {
length(&coder->rc, &coder->rep_len_encoder, pos_state, len,
coder->fast_mode);
update_long_rep(coder->state);
}
}
//////////
// Main //
//////////
static void
encode_symbol(lzma_lzma1_encoder *coder, lzma_mf *mf,
uint32_t back, uint32_t len, uint32_t position)
{
const uint32_t pos_state = position & coder->pos_mask;
if (back == UINT32_MAX) {
// Literal i.e. eight-bit byte
assert(len == 1);
rc_bit(&coder->rc,
&coder->is_match[coder->state][pos_state], 0);
literal(coder, mf, position);
} else {
// Some type of match
rc_bit(&coder->rc,
&coder->is_match[coder->state][pos_state], 1);
if (back < REPS) {
// It's a repeated match i.e. the same distance
// has been used earlier.
rc_bit(&coder->rc, &coder->is_rep[coder->state], 1);
rep_match(coder, pos_state, back, len);
} else {
// Normal match
rc_bit(&coder->rc, &coder->is_rep[coder->state], 0);
match(coder, pos_state, back - REPS, len);
}
}
assert(mf->read_ahead >= len);
mf->read_ahead -= len;
}
static bool
encode_init(lzma_lzma1_encoder *coder, lzma_mf *mf)
{
assert(mf_position(mf) == 0);
assert(coder->uncomp_size == 0);
if (mf->read_pos == mf->read_limit) {
if (mf->action == LZMA_RUN)
return false; // We cannot do anything.
// We are finishing (we cannot get here when flushing).
assert(mf->write_pos == mf->read_pos);
assert(mf->action == LZMA_FINISH);
} else {
// Do the actual initialization. The first LZMA symbol must
// always be a literal.
mf_skip(mf, 1);
mf->read_ahead = 0;
rc_bit(&coder->rc, &coder->is_match[0][0], 0);
rc_bittree(&coder->rc, coder->literal[0], 8, mf->buffer[0]);
++coder->uncomp_size;
}
// Initialization is done (except if empty file).
coder->is_initialized = true;
return true;
}
static void
encode_eopm(lzma_lzma1_encoder *coder, uint32_t position)
{
const uint32_t pos_state = position & coder->pos_mask;
rc_bit(&coder->rc, &coder->is_match[coder->state][pos_state], 1);
rc_bit(&coder->rc, &coder->is_rep[coder->state], 0);
match(coder, pos_state, UINT32_MAX, MATCH_LEN_MIN);
}
/// Number of bytes that a single encoding loop in lzma_lzma_encode() can
/// consume from the dictionary. This limit comes from lzma_lzma_optimum()
/// and may need to be updated if that function is significantly modified.
#define LOOP_INPUT_MAX (OPTS + 1)
extern lzma_ret
lzma_lzma_encode(lzma_lzma1_encoder *restrict coder, lzma_mf *restrict mf,
uint8_t *restrict out, size_t *restrict out_pos,
size_t out_size, uint32_t limit)
{
// Initialize the stream if no data has been encoded yet.
if (!coder->is_initialized && !encode_init(coder, mf))
return LZMA_OK;
// Encode pending output bytes from the range encoder.
// At the start of the stream, encode_init() encodes one literal.
// Later there can be pending output only with LZMA1 because LZMA2
// ensures that there is always enough output space. Thus when using
// LZMA2, rc_encode() calls in this function will always return false.
if (rc_encode(&coder->rc, out, out_pos, out_size)) {
// We don't get here with LZMA2.
assert(limit == UINT32_MAX);
return LZMA_OK;
}
// If the range encoder was flushed in an earlier call to this
// function but there wasn't enough output buffer space, those
// bytes would have now been encoded by the above rc_encode() call
// and the stream has now been finished. This can only happen with
// LZMA1 as LZMA2 always provides enough output buffer space.
if (coder->is_flushed) {
assert(limit == UINT32_MAX);
return LZMA_STREAM_END;
}
while (true) {
// With LZMA2 we need to take care that compressed size of
// a chunk doesn't get too big.
// FIXME? Check if this could be improved.
if (limit != UINT32_MAX
&& (mf->read_pos - mf->read_ahead >= limit
|| *out_pos + rc_pending(&coder->rc)
>= LZMA2_CHUNK_MAX
- LOOP_INPUT_MAX))
break;
// Check that there is some input to process.
if (mf->read_pos >= mf->read_limit) {
if (mf->action == LZMA_RUN)
return LZMA_OK;
if (mf->read_ahead == 0)
break;
}
// Get optimal match (repeat position and length).
// Value ranges for pos:
// - [0, REPS): repeated match
// - [REPS, UINT32_MAX):
// match at (pos - REPS)
// - UINT32_MAX: not a match but a literal
// Value ranges for len:
// - [MATCH_LEN_MIN, MATCH_LEN_MAX]
uint32_t len;
uint32_t back;
if (coder->fast_mode)
lzma_lzma_optimum_fast(coder, mf, &back, &len);
else
lzma_lzma_optimum_normal(coder, mf, &back, &len,
(uint32_t)(coder->uncomp_size));
encode_symbol(coder, mf, back, len,
(uint32_t)(coder->uncomp_size));
// If output size limiting is active (out_limit != 0), check
// if encoding this LZMA symbol would make the output size
// exceed the specified limit.
if (coder->out_limit != 0 && rc_encode_dummy(
&coder->rc, coder->out_limit)) {
// The most recent LZMA symbol would make the output
// too big. Throw it away.
rc_forget(&coder->rc);
// FIXME: Tell the LZ layer to not read more input as
// it would be waste of time. This doesn't matter if
// output-size-limited encoding is done with a single
// call though.
break;
}
// This symbol will be encoded so update the uncompressed size.
coder->uncomp_size += len;
// Encode the LZMA symbol.
if (rc_encode(&coder->rc, out, out_pos, out_size)) {
// Once again, this can only happen with LZMA1.
assert(limit == UINT32_MAX);
return LZMA_OK;
}
}
// Make the uncompressed size available to the application.
if (coder->uncomp_size_ptr != NULL)
*coder->uncomp_size_ptr = coder->uncomp_size;
// LZMA2 doesn't use EOPM at LZMA level.
//
// Plain LZMA streams without EOPM aren't supported except when
// output size limiting is enabled.
if (coder->use_eopm)
encode_eopm(coder, (uint32_t)(coder->uncomp_size));
// Flush the remaining bytes from the range encoder.
rc_flush(&coder->rc);
// Copy the remaining bytes to the output buffer. If there
// isn't enough output space, we will copy out the remaining
// bytes on the next call to this function.
if (rc_encode(&coder->rc, out, out_pos, out_size)) {
// This cannot happen with LZMA2.
assert(limit == UINT32_MAX);
coder->is_flushed = true;
return LZMA_OK;
}
return LZMA_STREAM_END;
}
static lzma_ret
lzma_encode(void *coder, lzma_mf *restrict mf,
uint8_t *restrict out, size_t *restrict out_pos,
size_t out_size)
{
// Plain LZMA has no support for sync-flushing.
if (unlikely(mf->action == LZMA_SYNC_FLUSH))
return LZMA_OPTIONS_ERROR;
return lzma_lzma_encode(coder, mf, out, out_pos, out_size, UINT32_MAX);
}
static lzma_ret
lzma_lzma_set_out_limit(
void *coder_ptr, uint64_t *uncomp_size, uint64_t out_limit)
{
// Minimum output size is 5 bytes but that cannot hold any output
// so we use 6 bytes.
if (out_limit < 6)
return LZMA_BUF_ERROR;
lzma_lzma1_encoder *coder = coder_ptr;
coder->out_limit = out_limit;
coder->uncomp_size_ptr = uncomp_size;
coder->use_eopm = false;
return LZMA_OK;
}
////////////////////
// Initialization //
////////////////////
static bool
is_options_valid(const lzma_options_lzma *options)
{
// Validate some of the options. LZ encoder validates nice_len too
// but we need a valid value here earlier.
return is_lclppb_valid(options)
&& options->nice_len >= MATCH_LEN_MIN
&& options->nice_len <= MATCH_LEN_MAX
&& (options->mode == LZMA_MODE_FAST
|| options->mode == LZMA_MODE_NORMAL);
}
static void
set_lz_options(lzma_lz_options *lz_options, const lzma_options_lzma *options)
{
// LZ encoder initialization does the validation for these so we
// don't need to validate here.
lz_options->before_size = OPTS;
lz_options->dict_size = options->dict_size;
lz_options->after_size = LOOP_INPUT_MAX;
lz_options->match_len_max = MATCH_LEN_MAX;
lz_options->nice_len = my_max(mf_get_hash_bytes(options->mf),
options->nice_len);
lz_options->match_finder = options->mf;
lz_options->depth = options->depth;
lz_options->preset_dict = options->preset_dict;
lz_options->preset_dict_size = options->preset_dict_size;
return;
}
static void
length_encoder_reset(lzma_length_encoder *lencoder,
const uint32_t num_pos_states, const bool fast_mode)
{
bit_reset(lencoder->choice);
bit_reset(lencoder->choice2);
for (size_t pos_state = 0; pos_state < num_pos_states; ++pos_state) {
bittree_reset(lencoder->low[pos_state], LEN_LOW_BITS);
bittree_reset(lencoder->mid[pos_state], LEN_MID_BITS);
}
bittree_reset(lencoder->high, LEN_HIGH_BITS);
if (!fast_mode)
for (uint32_t pos_state = 0; pos_state < num_pos_states;
++pos_state)
length_update_prices(lencoder, pos_state);
return;
}
extern lzma_ret
lzma_lzma_encoder_reset(lzma_lzma1_encoder *coder,
const lzma_options_lzma *options)
{
if (!is_options_valid(options))
return LZMA_OPTIONS_ERROR;
coder->pos_mask = (1U << options->pb) - 1;
coder->literal_context_bits = options->lc;
coder->literal_pos_mask = (1U << options->lp) - 1;
// Range coder
rc_reset(&coder->rc);
// State
coder->state = STATE_LIT_LIT;
for (size_t i = 0; i < REPS; ++i)
coder->reps[i] = 0;
literal_init(coder->literal, options->lc, options->lp);
// Bit encoders
for (size_t i = 0; i < STATES; ++i) {
for (size_t j = 0; j <= coder->pos_mask; ++j) {
bit_reset(coder->is_match[i][j]);
bit_reset(coder->is_rep0_long[i][j]);
}
bit_reset(coder->is_rep[i]);
bit_reset(coder->is_rep0[i]);
bit_reset(coder->is_rep1[i]);
bit_reset(coder->is_rep2[i]);
}
for (size_t i = 0; i < FULL_DISTANCES - DIST_MODEL_END; ++i)
bit_reset(coder->dist_special[i]);
// Bit tree encoders
for (size_t i = 0; i < DIST_STATES; ++i)
bittree_reset(coder->dist_slot[i], DIST_SLOT_BITS);
bittree_reset(coder->dist_align, ALIGN_BITS);
// Length encoders
length_encoder_reset(&coder->match_len_encoder,
1U << options->pb, coder->fast_mode);
length_encoder_reset(&coder->rep_len_encoder,
1U << options->pb, coder->fast_mode);
// Price counts are incremented every time appropriate probabilities
// are changed. price counts are set to zero when the price tables
// are updated, which is done when the appropriate price counts have
// big enough value, and lzma_mf.read_ahead == 0 which happens at
// least every OPTS (a few thousand) possible price count increments.
//
// By resetting price counts to UINT32_MAX / 2, we make sure that the
// price tables will be initialized before they will be used (since
// the value is definitely big enough), and that it is OK to increment
// price counts without risk of integer overflow (since UINT32_MAX / 2
// is small enough). The current code doesn't increment price counts
// before initializing price tables, but it maybe done in future if
// we add support for saving the state between LZMA2 chunks.
coder->match_price_count = UINT32_MAX / 2;
coder->align_price_count = UINT32_MAX / 2;
coder->opts_end_index = 0;
coder->opts_current_index = 0;
return LZMA_OK;
}
extern lzma_ret
lzma_lzma_encoder_create(void **coder_ptr, const lzma_allocator *allocator,
lzma_vli id, const lzma_options_lzma *options,
lzma_lz_options *lz_options)
{
assert(id == LZMA_FILTER_LZMA1 || id == LZMA_FILTER_LZMA1EXT
|| id == LZMA_FILTER_LZMA2);
// Allocate lzma_lzma1_encoder if it wasn't already allocated.
if (*coder_ptr == NULL) {
*coder_ptr = lzma_alloc(sizeof(lzma_lzma1_encoder), allocator);
if (*coder_ptr == NULL)
return LZMA_MEM_ERROR;
}
lzma_lzma1_encoder *coder = *coder_ptr;
// Set compression mode. Note that we haven't validated the options
// yet. Invalid options will get rejected by lzma_lzma_encoder_reset()
// call at the end of this function.
switch (options->mode) {
case LZMA_MODE_FAST:
coder->fast_mode = true;
break;
case LZMA_MODE_NORMAL: {
coder->fast_mode = false;
// Set dist_table_size.
// Round the dictionary size up to next 2^n.
//
// Currently the maximum encoder dictionary size
// is 1.5 GiB due to lz_encoder.c and here we need
// to be below 2 GiB to make the rounded up value
// fit in an uint32_t and avoid an infinite while-loop
// (and undefined behavior due to a too large shift).
// So do the same check as in LZ encoder,
// limiting to 1.5 GiB.
if (options->dict_size > (UINT32_C(1) << 30)
+ (UINT32_C(1) << 29))
return LZMA_OPTIONS_ERROR;
uint32_t log_size = 0;
while ((UINT32_C(1) << log_size) < options->dict_size)
++log_size;
coder->dist_table_size = log_size * 2;
// Length encoders' price table size
const uint32_t nice_len = my_max(
mf_get_hash_bytes(options->mf),
options->nice_len);
coder->match_len_encoder.table_size
= nice_len + 1 - MATCH_LEN_MIN;
coder->rep_len_encoder.table_size
= nice_len + 1 - MATCH_LEN_MIN;
break;
}
default:
return LZMA_OPTIONS_ERROR;
}
// We don't need to write the first byte as literal if there is
// a non-empty preset dictionary. encode_init() wouldn't even work
// if there is a non-empty preset dictionary, because encode_init()
// assumes that position is zero and previous byte is also zero.
coder->is_initialized = options->preset_dict != NULL
&& options->preset_dict_size > 0;
coder->is_flushed = false;
coder->uncomp_size = 0;
coder->uncomp_size_ptr = NULL;
// Output size limiting is disabled by default.
coder->out_limit = 0;
// Determine if end marker is wanted:
// - It is never used with LZMA2.
// - It is always used with LZMA_FILTER_LZMA1 (unless
// lzma_lzma_set_out_limit() is called later).
// - LZMA_FILTER_LZMA1EXT has a flag for it in the options.
coder->use_eopm = (id == LZMA_FILTER_LZMA1);
if (id == LZMA_FILTER_LZMA1EXT) {
// Check if unsupported flags are present.
if (options->ext_flags & ~LZMA_LZMA1EXT_ALLOW_EOPM)
return LZMA_OPTIONS_ERROR;
coder->use_eopm = (options->ext_flags
& LZMA_LZMA1EXT_ALLOW_EOPM) != 0;
// TODO? As long as there are no filters that change the size
// of the data, it is enough to look at lzma_stream.total_in
// after encoding has been finished to know the uncompressed
// size of the LZMA1 stream. But in the future there could be
// filters that change the size of the data and then total_in
// doesn't work as the LZMA1 stream size might be different
// due to another filter in the chain. The problem is simple
// to solve: Add another flag to ext_flags and then set
// coder->uncomp_size_ptr to the address stored in
// lzma_options_lzma.reserved_ptr2 (or _ptr1).
}
set_lz_options(lz_options, options);
return lzma_lzma_encoder_reset(coder, options);
}
static lzma_ret
lzma_encoder_init(lzma_lz_encoder *lz, const lzma_allocator *allocator,
lzma_vli id, const void *options, lzma_lz_options *lz_options)
{
if (options == NULL)
return LZMA_PROG_ERROR;
lz->code = &lzma_encode;
lz->set_out_limit = &lzma_lzma_set_out_limit;
return lzma_lzma_encoder_create(
&lz->coder, allocator, id, options, lz_options);
}
extern lzma_ret
lzma_lzma_encoder_init(lzma_next_coder *next, const lzma_allocator *allocator,
const lzma_filter_info *filters)
{
return lzma_lz_encoder_init(
next, allocator, filters, &lzma_encoder_init);
}
extern uint64_t
lzma_lzma_encoder_memusage(const void *options)
{
if (!is_options_valid(options))
return UINT64_MAX;
lzma_lz_options lz_options;
set_lz_options(&lz_options, options);
const uint64_t lz_memusage = lzma_lz_encoder_memusage(&lz_options);
if (lz_memusage == UINT64_MAX)
return UINT64_MAX;
return (uint64_t)(sizeof(lzma_lzma1_encoder)) + lz_memusage;
}
extern bool
lzma_lzma_lclppb_encode(const lzma_options_lzma *options, uint8_t *byte)
{
if (!is_lclppb_valid(options))
return true;
*byte = (options->pb * 5 + options->lp) * 9 + options->lc;
assert(*byte <= (4 * 5 + 4) * 9 + 8);
return false;
}
#ifdef HAVE_ENCODER_LZMA1
extern lzma_ret
lzma_lzma_props_encode(const void *options, uint8_t *out)
{
if (options == NULL)
return LZMA_PROG_ERROR;
const lzma_options_lzma *const opt = options;
if (lzma_lzma_lclppb_encode(opt, out))
return LZMA_PROG_ERROR;
write32le(out + 1, opt->dict_size);
return LZMA_OK;
}
#endif
extern LZMA_API(lzma_bool)
lzma_mode_is_supported(lzma_mode mode)
{
return mode == LZMA_MODE_FAST || mode == LZMA_MODE_NORMAL;
}