Merging upstream version 0.5.

Signed-off-by: Daniel Baumann <mail@daniel-baumann.ch>

15 files changed, 64 insertions, 59 deletions

diff --git a/AUTHORS b/AUTHORS
index 576d9a1..6badf8e 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -1,5 +1,4 @@
 Lzlib was written by Antonio Diaz Diaz.
 
-Lzlib implements a simplified version of the LZMA algorithm. The
-original LZMA algorithm was designed by Igor Pavlov.
-
+Lzlib implements a simplified version of the LZMA algorithm.
+The original LZMA algorithm was designed by Igor Pavlov.
diff --git a/ChangeLog b/ChangeLog
index d8ac767..b510dc0 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2009-07-03  Antonio Diaz Diaz  <ant_diaz@teleline.es>
+
+	* Version 0.5 released.
+	* Decompression speed has been improved.
+	* main.cc (signal_handler): Declared as `extern "C"'.
+
 2009-06-03  Antonio Diaz Diaz  <ant_diaz@teleline.es>
 
 	* Version 0.4 released.
diff --git a/Makefile.in b/Makefile.in
index 8c5f93c..1d05051 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -4,6 +4,7 @@ AR = ar
 INSTALL = install
 INSTALL_PROGRAM = $(INSTALL) -p -m 755
 INSTALL_DATA = $(INSTALL) -p -m 644
+INSTALL_DIR = $(INSTALL) -d -m 755
 LDCONFIG = ldconfig
 SHELL = /bin/sh
 
@@ -79,8 +80,8 @@ check : all $(VPATH)/testsuite/check.sh
 	@$(VPATH)/testsuite/check.sh $(VPATH)/testsuite
 
 install : all install-info
-	if test ! -d $(DESTDIR)$(includedir) ; then $(INSTALL) -d $(DESTDIR)$(includedir) ; fi
-	if test ! -d $(DESTDIR)$(libdir) ; then $(INSTALL) -d $(DESTDIR)$(libdir) ; fi
+	if test ! -d $(DESTDIR)$(includedir) ; then $(INSTALL_DIR) $(DESTDIR)$(includedir) ; fi
+	if test ! -d $(DESTDIR)$(libdir) ; then $(INSTALL_DIR) $(DESTDIR)$(libdir) ; fi
 	$(INSTALL_DATA) $(VPATH)/$(pkgname).h $(DESTDIR)$(includedir)/$(pkgname).h
 	$(INSTALL_DATA) ./$(libname).a $(DESTDIR)$(libdir)/$(libname).a
 	if test -n "$(progname_shared)" ; then \
@@ -94,10 +95,14 @@ install : all install-info
 	fi
 
 install-info :
-	if test ! -d $(DESTDIR)$(infodir) ; then $(INSTALL) -d $(DESTDIR)$(infodir) ; fi
+	if test ! -d $(DESTDIR)$(infodir) ; then $(INSTALL_DIR) $(DESTDIR)$(infodir) ; fi
 	$(INSTALL_DATA) $(VPATH)/doc/$(pkgname).info $(DESTDIR)$(infodir)/$(pkgname).info
 	-install-info --info-dir=$(DESTDIR)$(infodir) $(DESTDIR)$(infodir)/$(pkgname).info
 
+install-man :
+	if test ! -d $(DESTDIR)$(mandir)/man1 ; then $(INSTALL_DIR) $(DESTDIR)$(mandir)/man1 ; fi
+	$(INSTALL_DATA) $(VPATH)/doc/$(progname).1 $(DESTDIR)$(mandir)/man1/$(progname).1
+
 install-strip : all
 	$(MAKE) INSTALL_PROGRAM='$(INSTALL_PROGRAM) -s' install
 
@@ -108,9 +113,12 @@ uninstall : uninstall-info
 	-rm -f $(DESTDIR)$(libdir)/$(libname).so.$(pkgversion)
 
 uninstall-info :
-	-install-info --remove $(DESTDIR)$(infodir)/$(pkgname).info $(DESTDIR)$(infodir)/dir
+	-install-info --info-dir=$(DESTDIR)$(infodir) --remove $(DESTDIR)$(infodir)/$(pkgname).info
 	-rm -f $(DESTDIR)$(infodir)/$(pkgname).info
 
+uninstall-man :
+	-rm -f $(DESTDIR)$(mandir)/man1/$(progname).1
+
 dist :
 	ln -sf $(VPATH) $(DISTNAME)
 	tar -cvf $(DISTNAME).tar \
@@ -132,7 +140,7 @@ dist :
 	lzip -v -9 $(DISTNAME).tar
 
 clean :
-	-rm -f $(progname) $(progname)_profiled $(lib_objs) $(objs) *.a
+	-rm -f $(progname) $(progname)_profiled $(objs) $(lib_objs) *.a
 	-rm -f $(progname)_shared $(sh_lib_objs) *.so.$(pkgversion)
 
 distclean : clean
diff --git a/NEWS b/NEWS
index f97bcd8..7a7df99 100644
--- a/NEWS
+++ b/NEWS
@@ -1,10 +1,5 @@
-Changes in version 0.4:
-
-Partial flush of the compressed data has been implemented with the
-function LZ_compress_sync_flush.
-
-The function LZ_compress_write_size has been added.
+Changes in version 0.5:
 
 Decompression speed has been improved.
 
-The chapter "Buffering" has been added to the manual.
+A build problem with the Sun C++ compiler has been fixed.
diff --git a/configure b/configure
index ba2cad0..4255d2f 100755
--- a/configure
+++ b/configure
@@ -5,13 +5,13 @@
 # This configure script is free software: you have unlimited permission
 # to copy, distribute and modify it.
 #
-# Date of this version: 2009-06-03
+# Date of this version: 2009-07-03
 
 invocation_name=$0
 args=
 no_create=
 pkgname=lzlib
-pkgversion=0.4
+pkgversion=0.5
 soversion=0
 progname=minilzip
 progname_shared=${progname}_shared
diff --git a/decoder.cc b/decoder.cc
index f68edf9..b426b3e 100644
--- a/decoder.cc
+++ b/decoder.cc
@@ -130,12 +130,12 @@ int LZ_decoder::decode_member()
     const int pos_state = data_position() & pos_state_mask;
     if( range_decoder.decode_bit( bm_match[state()][pos_state] ) == 0 )
       {
+      const uint8_t prev_byte = get_byte( 0 );
       if( state.is_char() )
-        prev_byte = literal_decoder.decode( range_decoder, prev_byte );
+        put_byte( literal_decoder.decode( range_decoder, prev_byte ) );
       else
-        prev_byte = literal_decoder.decode_matched( range_decoder, prev_byte,
-                                                    get_byte( rep0 ) );
-      put_byte( prev_byte );
+        put_byte( literal_decoder.decode_matched( range_decoder, prev_byte,
+                                                  get_byte( rep0 ) ) );
       state.set_char();
       }
     else
@@ -209,7 +209,6 @@ int LZ_decoder::decode_member()
         state.set_match();
         }
       copy_block( rep0, len );
-      prev_byte = get_byte( 0 );
       }
     }
   }
diff --git a/decoder.h b/decoder.h
index 6484898..bc9297a 100644
--- a/decoder.h
+++ b/decoder.h
@@ -91,9 +91,9 @@ public:
     if( force ) reload_pending = true;
     if( reload_pending && available_bytes() >= 5 )
       {
+      reload_pending = false;
       code = 0;
       range = 0xFFFFFFFF;
-      reload_pending = false;
       for( int i = 0; i < 5; ++i ) code = (code << 8) | get_byte();
       }
     return !reload_pending;
@@ -216,11 +216,11 @@ class Literal_decoder
     { return ( prev_byte >> ( 8 - literal_context_bits ) ); }
 
 public:
-  uint8_t decode( Range_decoder & range_decoder, const int prev_byte )
+  uint8_t decode( Range_decoder & range_decoder, const uint8_t prev_byte )
     { return range_decoder.decode_tree( bm_literal[state(prev_byte)], 8 ); }
 
   uint8_t decode_matched( Range_decoder & range_decoder,
-                          const int prev_byte, const int match_byte )
+                          const uint8_t prev_byte, const uint8_t match_byte )
     { return range_decoder.decode_matched( bm_literal[state(prev_byte)], match_byte ); }
   };
 
@@ -233,12 +233,11 @@ class LZ_decoder : public Circular_buffer
   const int dictionary_size;
   uint32_t crc_;
   bool member_finished_;
-  unsigned int rep0;
-  unsigned int rep1;
-  unsigned int rep2;
+  unsigned int rep0;		// rep[0-3] latest four distances
+  unsigned int rep1;		// used for efficient coding of
+  unsigned int rep2;		// repeated distances
   unsigned int rep3;
   State state;
-  uint8_t prev_byte;
 
   Bit_model bm_match[State::states][pos_states];
   Bit_model bm_rep[State::states];
@@ -273,7 +272,7 @@ class LZ_decoder : public Circular_buffer
     {
     int i = put - distance - 1;
     if( i < 0 ) i += buffer_size;
-    if( len < buffer_size - std::max( put, i ) && len <= distance )
+    if( len < buffer_size - std::max( put, i ) && len <= std::abs( put - i ) )
       {
       crc32.update( crc_, buffer + i, len );
       std::memcpy( buffer + put, buffer + i, len );
@@ -303,9 +302,9 @@ public:
     rep1( 0 ),
     rep2( 0 ),
     rep3( 0 ),
-    prev_byte( 0 ),
     range_decoder( sizeof header, ibuf ),
-    literal_decoder() {}
+    literal_decoder()
+    { buffer[buffer_size-1] = 0; }	// prev_byte of first_byte
 
   bool enough_free_bytes() const throw()
     { return free_bytes() >= min_free_bytes; }
diff --git a/doc/lzlib.info b/doc/lzlib.info
index fff59c2..e688ea6 100644
--- a/doc/lzlib.info
+++ b/doc/lzlib.info
@@ -12,7 +12,7 @@ File: lzlib.info,  Node: Top,  Next: Introduction,  Up: (dir)
 Lzlib
 *****
 
-This manual is for Lzlib (version 0.4, 3 June 2009).
+This manual is for Lzlib (version 0.5, 3 July 2009).
 
 * Menu:
 
diff --git a/doc/lzlib.texinfo b/doc/lzlib.texinfo
index 044bd04..d0cdd2b 100644
--- a/doc/lzlib.texinfo
+++ b/doc/lzlib.texinfo
@@ -5,8 +5,8 @@
 @finalout
 @c %**end of header
 
-@set UPDATED 3 June 2009
-@set VERSION 0.4
+@set UPDATED 3 July 2009
+@set VERSION 0.5
 
 @dircategory Data Compression
 @direntry
diff --git a/encoder.cc b/encoder.cc
index cca154c..51ffb95 100644
--- a/encoder.cc
+++ b/encoder.cc
@@ -61,9 +61,8 @@ Matchfinder::Matchfinder( const int dict_size, const int len_limit )
   :
   partial_data_pos( 0 ),
   dictionary_size_( dict_size ),
-  after_size( max_num_trials + max_match_len ),
   buffer_size( ( 2 * std::max( 65536, dictionary_size_ ) ) +
-               max_num_trials + after_size ),
+               before_size + after_size ),
   buffer( new( std::nothrow ) uint8_t[buffer_size] ),
   pos( 0 ),
   cyclic_pos( 0 ),
@@ -105,7 +104,7 @@ bool Matchfinder::move_pos() throw()
     if( pos > stream_pos ) { pos = stream_pos; return false; }
     else
       {
-      const int offset = pos - dictionary_size_ - max_num_trials;
+      const int offset = pos - dictionary_size_ - before_size;
       const int size = stream_pos - offset;
       std::memmove( buffer, buffer + offset, size );
       partial_data_pos += offset;
@@ -457,7 +456,7 @@ int LZ_encoder::best_pair_sequence( const int reps[num_rep_distances],
   }
 
 
-     // Sync Flush mark => (dis == 0xFFFFFFFF, len == min_match_len+1)
+     // Sync Flush mark => (dis == 0xFFFFFFFF, len == min_match_len + 1)
 bool LZ_encoder::sync_flush()
   {
   if( member_finished_ || range_encoder.free_bytes() < max_marker_size )
@@ -505,7 +504,6 @@ LZ_encoder::LZ_encoder( Matchfinder & mf, const File_header & header,
   literal_encoder(),
   num_dis_slots( 2 * File_header::real_bits( matchfinder.dictionary_size() - 1 ) ),
   fill_counter( 0 ),
-  prev_byte( 0 ),
   member_finished_( false )
   {
   for( int i = 0; i < num_rep_distances; ++i ) rep_distances[i] = 0;
@@ -522,15 +520,15 @@ bool LZ_encoder::encode_member( const bool finish )
   if( range_encoder.member_position() >= member_size_limit )
     { if( full_flush() ) { member_finished_ = true; } return true; }
 
-  // copy first byte
+  // encode first byte
   if( matchfinder.data_position() == 0 && !matchfinder.finished() )
     {
     if( matchfinder.available_bytes() < 4 && !matchfinder.at_stream_end() )
       return true;
     range_encoder.encode_bit( bm_match[state()][0], 0 );
+    const uint8_t prev_byte = 0;
     const uint8_t cur_byte = matchfinder[0];
     literal_encoder.encode( range_encoder, prev_byte, cur_byte );
-    prev_byte = cur_byte;
     crc32.update( crc_, cur_byte );
     if( !move_pos( 1 ) ) return false;
     }
@@ -560,16 +558,16 @@ bool LZ_encoder::encode_member( const bool finish )
       range_encoder.encode_bit( bm_match[state()][pos_state], !bit );
       if( bit )
         {
+        const uint8_t prev_byte = matchfinder[-ahead-1];
         const uint8_t cur_byte = matchfinder[-ahead];
         if( state.is_char() )
           literal_encoder.encode( range_encoder, prev_byte, cur_byte );
         else
           {
-          const uint8_t match_byte = matchfinder[-rep_distances[0]-1-ahead];
+          const uint8_t match_byte = matchfinder[-ahead-rep_distances[0]-1];
           literal_encoder.encode_matched( range_encoder, prev_byte, match_byte, cur_byte );
           }
         state.set_char();
-        prev_byte = cur_byte;
         }
       else
         {
@@ -600,7 +598,6 @@ bool LZ_encoder::encode_member( const bool finish )
           encode_pair( dis - num_rep_distances, len, pos_state );
           state.set_match();
           }
-        prev_byte = matchfinder[len-1-ahead];
         }
       for( int j = 0; j < len; ++j )
         crc32.update( crc_, matchfinder[j-ahead] );
diff --git a/encoder.h b/encoder.h
index cdfd751..9583e4f 100644
--- a/encoder.h
+++ b/encoder.h
@@ -145,7 +145,11 @@ inline int price_matched( const Bit_model bm[], const int symbol,
 
 class Matchfinder
   {
-  enum { num_prev_positions4 = 1 << 20,
+  enum { // bytes to keep in buffer before dictionary
+         before_size = max_num_trials + 1,
+         // bytes to keep in buffer after pos
+         after_size = max_num_trials + max_match_len,
+         num_prev_positions4 = 1 << 20,
          num_prev_positions3 = 1 << 18,
          num_prev_positions2 = 1 << 16,
          num_prev_positions = num_prev_positions4 + num_prev_positions3 +
@@ -153,7 +157,6 @@ class Matchfinder
 
   long long partial_data_pos;
   const int dictionary_size_;	// bytes to keep in buffer before pos
-  const int after_size;		// bytes to keep in buffer after pos
   const int buffer_size;
   uint8_t * const buffer;
   int pos;
@@ -459,7 +462,6 @@ class LZ_encoder
   int align_price_count;
   int fill_counter;
   State state;
-  uint8_t prev_byte;
   bool member_finished_;
 
   void fill_align_prices() throw();
diff --git a/lzlib.cc b/lzlib.cc
index c64ed43..76212db 100644
--- a/lzlib.cc
+++ b/lzlib.cc
@@ -335,7 +335,7 @@ int LZ_decompress_read( void * const decoder, uint8_t * const buffer,
     if( d.ibuf->used_bytes() < 5 + (int)sizeof( File_header ) )
       {
       if( !d.ibuf->at_stream_end() || d.ibuf->finished() ) return 0;
-      d.ibuf->purge();
+      d.ibuf->purge();			// remove trailing garbage
       d.lz_errno = LZ_header_error;
       return -1;
       }
@@ -346,12 +346,12 @@ int LZ_decompress_read( void * const decoder, uint8_t * const buffer,
         header.dictionary_size() < min_dictionary_size ||
         header.dictionary_size() > max_dictionary_size )
       {
-      d.ibuf->purge();
+      d.ibuf->purge();			// remove trailing garbage
       d.lz_errno = LZ_header_error;
       return -1;
       }
     try { d.lz_decoder = new LZ_decoder( header, *d.ibuf ); }
-    catch( std::bad_alloc )
+    catch( std::bad_alloc )		// not enough free memory
       {
       d.ibuf->purge();
       d.lz_decoder = 0;
diff --git a/lzlib.h b/lzlib.h
index f2ef9d3..7dc9aab 100644
--- a/lzlib.h
+++ b/lzlib.h
@@ -29,7 +29,7 @@
 extern "C" {
 #endif
 
-const char * const LZ_version_string = "0.4";
+const char * const LZ_version_string = "0.5";
 
 enum { min_dictionary_bits = 12,
        min_dictionary_size = 1 << min_dictionary_bits,
diff --git a/main.cc b/main.cc
index d9cf922..11ee21f 100644
--- a/main.cc
+++ b/main.cc
@@ -558,8 +558,8 @@ int decompress( const int inhandle, const Pretty_print & pp,
       const LZ_errno lz_errno = LZ_decompress_errno( decoder );
       if( lz_errno == LZ_header_error )
         {
-        if( LZ_decompress_total_out_size( decoder ) > 0 )  // trailing garbage
-          break;
+        if( LZ_decompress_total_out_size( decoder ) > 0 )
+          break;				// trailing garbage
         pp( "error reading member header" );
         return 1;
         }
@@ -596,7 +596,7 @@ int decompress( const int inhandle, const Pretty_print & pp,
   }
 
 
-void signal_handler( const int ) throw()
+extern "C" void signal_handler( int ) throw()
   {
   show_error( "Control-C or similar caught, quitting." );
   cleanup_and_fail( 0 );
diff --git a/testsuite/check.sh b/testsuite/check.sh
index 4431864..2ce2ca5 100755
--- a/testsuite/check.sh
+++ b/testsuite/check.sh
@@ -28,7 +28,7 @@ fail=0
 "${LZIP}" -cd "${testdir}"/COPYING.lz > copy || fail=1
 cmp in copy || fail=1
 
-for i in 1 2 3 4 5 6 7 8 9; do
+for i in s4096 1 2 3 4 5 6 7 8 9; do
 	"${LZIP}" -k -$i in || fail=1
 	mv -f in.lz copy.lz || fail=1
 	echo -n "garbage" >> copy.lz || fail=1
@@ -37,7 +37,7 @@ for i in 1 2 3 4 5 6 7 8 9; do
 	echo -n .
 done
 
-for i in 1 2 3 4 5 6 7 8 9; do
+for i in s4096 1 2 3 4 5 6 7 8 9; do
 	"${LZIP}" -c -$i in > out || fail=1
 	echo -n "g" >> out || fail=1
 	"${LZIP}" -cd out > copy || fail=1
@@ -45,14 +45,14 @@ for i in 1 2 3 4 5 6 7 8 9; do
 	echo -n .
 done
 
-for i in 1 2 3 4 5 6 7 8 9; do
+for i in s4096 1 2 3 4 5 6 7 8 9; do
 	"${LZIP}" -c -$i < in > out || fail=1
 	"${LZIP}" -d < out > copy || fail=1
 	cmp in copy || fail=1
 	echo -n .
 done
 
-for i in 1 2 3 4 5 6 7 8 9; do
+for i in s4096 1 2 3 4 5 6 7 8 9; do
 	"${LZIP}" -f -$i -o out < in || fail=1
 	"${LZIP}" -df -o copy < out.lz || fail=1
 	cmp in copy || fail=1