Added initial support for preset dictionary for raw LZMA1

and LZMA2. It is not supported by the .xz format or the xz
command line tool yet.

5 files changed, 28 insertions, 14 deletions

diff --git a/src/liblzma/lzma/lzma2_decoder.c b/src/liblzma/lzma/lzma2_decoder.c
index ff90803b..fe925a59 100644
--- a/src/liblzma/lzma/lzma2_decoder.c
+++ b/src/liblzma/lzma/lzma2_decoder.c
@@ -230,7 +230,7 @@ lzma2_decoder_end(lzma_coder *coder, lzma_allocator *allocator)
 
 static lzma_ret
 lzma2_decoder_init(lzma_lz_decoder *lz, lzma_allocator *allocator,
-		const void *options, size_t *dict_size)
+		const void *opt, lzma_lz_options *lz_options)
 {
 	if (lz->coder == NULL) {
 		lz->coder = lzma_alloc(sizeof(lzma_coder), allocator);
@@ -243,12 +243,15 @@ lzma2_decoder_init(lzma_lz_decoder *lz, lzma_allocator *allocator,
 		lz->coder->lzma = LZMA_LZ_DECODER_INIT;
 	}
 
+	const lzma_options_lzma *options = opt;
+
 	lz->coder->sequence = SEQ_CONTROL;
 	lz->coder->need_properties = true;
-	lz->coder->need_dictionary_reset = true;
+	lz->coder->need_dictionary_reset = options->preset_dict == NULL
+			|| options->preset_dict_size == 0;
 
 	return lzma_lzma_decoder_create(&lz->coder->lzma,
-			allocator, options, dict_size);
+			allocator, options, lz_options);
 }
 
 
diff --git a/src/liblzma/lzma/lzma2_encoder.c b/src/liblzma/lzma/lzma2_encoder.c
index 81b6f973..cc676d5e 100644
--- a/src/liblzma/lzma/lzma2_encoder.c
+++ b/src/liblzma/lzma/lzma2_encoder.c
@@ -318,15 +318,17 @@ lzma2_encoder_init(lzma_lz_encoder *lz, lzma_allocator *allocator,
 		lz->coder->lzma = NULL;
 	}
 
-	lz->coder->sequence = SEQ_INIT;
-	lz->coder->need_properties = true;
-	lz->coder->need_state_reset = false;
-	lz->coder->need_dictionary_reset = true;
-
 	lz->coder->opt_cur = *(const lzma_options_lzma *)(options);
 	lz->coder->opt_new = lz->coder->opt_cur.persistent
 				? options : NULL;
 
+	lz->coder->sequence = SEQ_INIT;
+	lz->coder->need_properties = true;
+	lz->coder->need_state_reset = false;
+	lz->coder->need_dictionary_reset
+			= lz->coder->opt_cur.preset_dict == NULL
+			|| lz->coder->opt_cur.preset_dict_size == 0;
+
 	// Initialize LZMA encoder
 	return_if_error(lzma_lzma_encoder_create(&lz->coder->lzma, allocator,
 			&lz->coder->opt_cur, lz_options));
diff --git a/src/liblzma/lzma/lzma_decoder.c b/src/liblzma/lzma/lzma_decoder.c
index 03e3251a..80a9fd6c 100644
--- a/src/liblzma/lzma/lzma_decoder.c
+++ b/src/liblzma/lzma/lzma_decoder.c
@@ -941,7 +941,7 @@ lzma_decoder_reset(lzma_coder *coder, const void *opt)
 
 extern lzma_ret
 lzma_lzma_decoder_create(lzma_lz_decoder *lz, lzma_allocator *allocator,
-		const void *opt, size_t *dict_size)
+		const void *opt, lzma_lz_options *lz_options)
 {
 	if (lz->coder == NULL) {
 		lz->coder = lzma_alloc(sizeof(lzma_coder), allocator);
@@ -956,7 +956,9 @@ lzma_lzma_decoder_create(lzma_lz_decoder *lz, lzma_allocator *allocator,
 	// All dictionary sizes are OK here. LZ decoder will take care of
 	// the special cases.
 	const lzma_options_lzma *options = opt;
-	*dict_size = options->dict_size;
+	lz_options->dict_size = options->dict_size;
+	lz_options->preset_dict = options->preset_dict;
+	lz_options->preset_dict_size = options->preset_dict_size;
 
 	return LZMA_OK;
 }
@@ -967,13 +969,13 @@ lzma_lzma_decoder_create(lzma_lz_decoder *lz, lzma_allocator *allocator,
 /// the LZ initialization).
 static lzma_ret
 lzma_decoder_init(lzma_lz_decoder *lz, lzma_allocator *allocator,
-		const void *options, size_t *dict_size)
+		const void *options, lzma_lz_options *lz_options)
 {
 	if (!is_lclppb_valid(options))
 		return LZMA_PROG_ERROR;
 
 	return_if_error(lzma_lzma_decoder_create(
-			lz, allocator, options, dict_size));
+			lz, allocator, options, lz_options));
 
 	lzma_decoder_reset(lz->coder, options);
 	lzma_decoder_uncompressed(lz->coder, LZMA_VLI_UNKNOWN);
diff --git a/src/liblzma/lzma/lzma_decoder.h b/src/liblzma/lzma/lzma_decoder.h
index 133d2608..15844052 100644
--- a/src/liblzma/lzma/lzma_decoder.h
+++ b/src/liblzma/lzma/lzma_decoder.h
@@ -48,7 +48,7 @@ extern bool lzma_lzma_lclppb_decode(
 /// LZMA2 decoders.
 extern lzma_ret lzma_lzma_decoder_create(
 		lzma_lz_decoder *lz, lzma_allocator *allocator,
-		const void *opt, size_t *dict_size);
+		const void *opt, lzma_lz_options *lz_options);
 
 /// Gets memory usage without validating lc/lp/pb. This is used by LZMA2
 /// decoder, because raw LZMA2 decoding doesn't need lc/lp/pb.
diff --git a/src/liblzma/lzma/lzma_encoder.c b/src/liblzma/lzma/lzma_encoder.c
index 79bb8f9b..cf7637d2 100644
--- a/src/liblzma/lzma/lzma_encoder.c
+++ b/src/liblzma/lzma/lzma_encoder.c
@@ -274,6 +274,8 @@ encode_symbol(lzma_coder *coder, lzma_mf *mf,
 static bool
 encode_init(lzma_coder *coder, lzma_mf *mf)
 {
+	assert(mf_position(mf) == 0);
+
 	if (mf->read_pos == mf->read_limit) {
 		if (mf->action == LZMA_RUN)
 			return false; // We cannot do anything.
@@ -594,7 +596,12 @@ lzma_lzma_encoder_create(lzma_coder **coder_ptr, lzma_allocator *allocator,
 			return LZMA_OPTIONS_ERROR;
 	}
 
-	coder->is_initialized = false;
+	// 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;
 
 	set_lz_options(lz_options, options);