Merging upstream version 0.3.

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

1 files changed, 51 insertions, 394 deletions

diff --git a/main.cc b/main.cc
index 18ed642..f3f12e0 100644
--- a/main.cc
+++ b/main.cc
@@ -25,20 +25,17 @@
 #define _FILE_OFFSET_BITS 64
 
 #include <algorithm>
-#include <cassert>
 #include <cerrno>
 #include <climits>
 #include <csignal>
-#include <cstdarg>
-#include <cstddef>
 #include <cstdio>
 #include <cstdlib>
 #include <cstring>
 #include <string>
 #include <vector>
 #include <fcntl.h>
-#include <stdint.h>
 #include <pthread.h>
+#include <stdint.h>
 #include <unistd.h>
 #include <utime.h>
 #include <sys/stat.h>
@@ -49,7 +46,6 @@
 #endif
 
 #include "arg_parser.h"
-#include "main.h"
 #include "plzip.h"
 
 #ifndef LLONG_MAX
@@ -72,6 +68,12 @@ const char * const Program_name    = "Plzip";
 const char * const program_name    = "plzip";
 const char * const program_year    = "2010";
 
+#ifdef O_BINARY
+const int o_binary = O_BINARY;
+#else
+const int o_binary = 0;
+#endif
+
 struct { const char * from; const char * to; } const known_extensions[] = {
   { ".lz",  ""     },
   { ".tlz", ".tar" },
@@ -88,6 +90,8 @@ enum Mode { m_compress = 0, m_decompress, m_test };
 std::string output_filename;
 int outhandle = -1;
 bool delete_output_on_interrupt = false;
+pthread_t main_thread;
+pid_t main_thread_pid;
 
 class Pretty_print
   {
@@ -132,6 +136,7 @@ void show_help() throw()
   std::printf( "  -h, --help                 display this help and exit\n" );
   std::printf( "  -V, --version              output version information and exit\n" );
 //  std::printf( "  -b, --member-size=<n>      set member size limit in bytes\n" );
+  std::printf( "  -B, --data-size=<n>        set input data block size in bytes\n" );
   std::printf( "  -c, --stdout               send output to standard output\n" );
   std::printf( "  -d, --decompress           decompress\n" );
   std::printf( "  -f, --force                overwrite existing output files\n" );
@@ -267,7 +272,7 @@ int open_instream( const std::string & name, struct stat * in_statsp,
     }
   else
     {
-    inhandle = open( name.c_str(), O_RDONLY );
+    inhandle = open( name.c_str(), O_RDONLY | o_binary );
     if( inhandle < 0 )
       {
       if( verbosity >= 0 )
@@ -324,9 +329,11 @@ void set_d_outname( const std::string & name, const int i ) throw()
 bool open_outstream( const bool force ) throw()
   {
   if( force )
-    outhandle = open( output_filename.c_str(), O_CREAT | O_TRUNC | O_WRONLY,
+    outhandle = open( output_filename.c_str(),
+                      O_CREAT | O_TRUNC | O_WRONLY | o_binary,
                       S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH );
-  else outhandle = open( output_filename.c_str(), O_CREAT | O_EXCL | O_WRONLY,
+  else outhandle = open( output_filename.c_str(),
+                         O_CREAT | O_EXCL | O_WRONLY | o_binary,
                          S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH );
   if( outhandle < 0 )
     {
@@ -412,14 +419,13 @@ int do_decompress( LZ_Decoder * const decoder, const int inhandle,
   const int in_buffer_size = 65536, out_buffer_size = 8 * in_buffer_size;
   uint8_t in_buffer[in_buffer_size], out_buffer[out_buffer_size];
 
-  if( verbosity >= 1 ) pp();
   while( true )
     {
     int in_size = std::min( LZ_decompress_write_size( decoder ), in_buffer_size );
     if( in_size > 0 )
       {
       const int max_in_size = in_size;
-      in_size = readblock( inhandle, (char *)in_buffer, max_in_size );
+      in_size = readblock( inhandle, in_buffer, max_in_size );
       if( in_size != max_in_size && errno )
         { pp(); show_error( "read error", errno ); return 1; }
       if( in_size == 0 ) LZ_decompress_finish( decoder );
@@ -458,7 +464,7 @@ int do_decompress( LZ_Decoder * const decoder, const int inhandle,
       }
     else if( out_size > 0 && outhandle >= 0 )
       {
-      const int wr = writeblock( outhandle, (char *)out_buffer, out_size );
+      const int wr = writeblock( outhandle, out_buffer, out_size );
       if( wr != out_size )
         { pp(); show_error( "write error", errno ); return 1; }
       }
@@ -490,9 +496,12 @@ int decompress( const int inhandle, const Pretty_print & pp,
   }
 
 
-extern "C" void signal_handler( int ) throw()
+extern "C" void signal_handler( int sig ) throw()
   {
-  show_error( "Control-C or similar caught, quitting." );
+  if( !pthread_equal( pthread_self(), main_thread ) )
+    kill( main_thread_pid, sig );
+  if( sig != SIGUSR1 )
+    show_error( "Control-C or similar caught, quitting." );
   cleanup_and_fail( 1 );
   }
 
@@ -502,6 +511,7 @@ void set_signals() throw()
   signal( SIGHUP, signal_handler );
   signal( SIGINT, signal_handler );
   signal( SIGTERM, signal_handler );
+  signal( SIGUSR1, signal_handler );
   }
 
 } // end namespace
@@ -551,372 +561,16 @@ void internal_error( const char * msg )
   }
 
 
-/* Private stuff needed by fatal(). */
-static pthread_t main_thread;
-
-static pid_t pid;
-
-
-/* Public utility variables and functions. */
-
-/*
-  This can be called from any thread, main thread or sub-threads alike, since
-  they all call common helper functions that call fatal() in case of an error.
-*/
-void fatal()
-{
-  if( pthread_equal(pthread_self(), main_thread) )
-    cleanup_and_fail( 1 );
-  else
-    {
-    if( 0 == kill(pid, SIGUSR1) )
-      pthread_exit(0);
-    }
-  _exit( 1 );
-}
-
-
-void
-fail(const char *fmt, int err, ...)
-{
-  va_list args;
-
-  /* Locking stderr should also protect strerror(). */
-  flockfile(stderr);
-  (void)fprintf(stderr, "%s: ", program_name);
-
-  va_start(args, err);
-  (void)vfprintf(stderr, fmt, args);
-  va_end(args);
-
-  (void)fprintf(stderr, ": %s\n", strerror(err));
-  funlockfile(stderr);
-  /* Stream stderr is never fully buffered originally. */
-  fatal();
-}
-
-
-void
-xinit(Cond *cond)
-{
-  pthread_mutexattr_t attr;
-
-  int ret = pthread_mutexattr_init(&attr);
-  if( ret != 0 ) {
-    fail("pthread_mutexattr_init()", ret);
-  }
-
-  ret = pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK);
-  if( ret != 0 ) {
-    fail("pthread_mutexattr_settype()", ret);
-  }
-
-  ret = pthread_mutex_init(&cond->lock, &attr);
-  if( ret != 0 ) {
-    fail("pthread_mutex_init()", ret);
-  }
-
-  ret = pthread_mutexattr_destroy(&attr);
-  if( ret != 0 ) {
-    fail("pthread_mutexattr_destroy()", ret);
-  }
-
-  ret = pthread_cond_init(&cond->cond, 0);
-  if( ret != 0 ) {
-    fail("pthread_cond_init()", ret);
-  }
-
-  cond->ccount = 0;
-  cond->wcount = 0;
-}
-
-
-void
-xdestroy(Cond *cond)
-{
-  int ret = pthread_cond_destroy(&cond->cond);
-  if( ret != 0 ) {
-    fail("pthread_cond_destroy()", ret);
-  }
-
-  ret = pthread_mutex_destroy(&cond->lock);
-  if( ret != 0 ) {
-    fail("pthread_mutex_destroy()", ret);
-  }
-}
-
-
-void
-xlock(Cond *cond)
-{
-  int ret = pthread_mutex_lock(&cond->lock);
-  if( ret != 0 ) {
-    fail("pthread_mutex_lock()", ret);
-  }
-}
-
-
-void
-xlock_pred(Cond *cond)
-{
-  xlock(cond);
-  ++cond->ccount;
-}
-
-
-void
-xunlock(Cond *cond)
-{
-  int ret = pthread_mutex_unlock(&cond->lock);
-  if( ret != 0 ) {
-    fail("pthread_mutex_unlock()", ret);
-  }
-}
-
-
-void
-xwait(Cond *cond)
-{
-  ++cond->wcount;
-  int ret = pthread_cond_wait(&cond->cond, &cond->lock);
-  if( ret != 0 ) {
-    fail("pthread_cond_wait()", ret);
-  }
-  ++cond->ccount;
-}
-
-
-void
-xsignal(Cond *cond)
-{
-  int ret = pthread_cond_signal(&cond->cond);
-  if( ret != 0 ) {
-    fail("pthread_cond_signal()", ret);
-  }
-}
-
-
-void
-xbroadcast(Cond *cond)
-{
-  int ret = pthread_cond_broadcast(&cond->cond);
-  if( ret != 0 ) {
-    fail("pthread_cond_broadcast()", ret);
-  }
-}
-
-
-void
-xcreate(pthread_t *thread, void *(*routine)(void *), void *arg)
-{
-  int ret = pthread_create(thread, 0, routine, arg);
-  if( ret != 0 ) {
-    fail("pthread_create()", ret);
-  }
-}
-
-
-void
-xjoin(pthread_t thread)
-{
-  int ret = pthread_join(thread, 0);
-  if( ret != 0 ) {
-    fail("pthread_join()", ret);
-  }
-}
-
-
-void
-xraise(int sig)
-{
-  if( -1 == kill(pid, sig) ) {
-    fail("kill()", errno);
-  }
-}
-
-
-/* Private stuff part 2. */
-
-
-static void
-xsigemptyset(sigset_t *set)
-{
-  if( -1 == sigemptyset(set) ) {
-    fail("sigemptyset()", errno);
-  }
-}
-
-
-static void
-xsigaddset(sigset_t *set, int signo)
-{
-  if( -1 == sigaddset(set, signo) ) {
-    fail("sigaddset()", errno);
-  }
-}
-
-
-static void
-xsigmask(int how, const sigset_t *set, sigset_t *oset)
-{
-  int ret = pthread_sigmask(how, set, oset);
-  if( ret != 0 ) {
-    fail("pthread_sigmask()", ret);
-  }
-}
-
-
-static void
-xsigaction(int sig, void (*handler)(int))
-{
-  struct sigaction act;
-
-  act.sa_handler = handler;
-  xsigemptyset(&act.sa_mask);
-  act.sa_flags = 0;
-
-  if( -1 == sigaction(sig, &act, 0) ) {
-    fail("sigaction()", errno);
-  }
-}
-
-
-enum Caught_sig { CS_INT = 1, CS_TERM, CS_USR1, CS_USR2 };
-
-static volatile sig_atomic_t caught_sig;
-
-
-extern "C" void sighandler( int sig )
-  {
-  /* sig_atomic_t is nowhere required to be able to hold signal values. */
-  switch( sig )
-    {
-    case SIGINT : caught_sig = CS_INT;  break;
-    case SIGTERM: caught_sig = CS_TERM; break;
-    case SIGUSR1: caught_sig = CS_USR1; break;
-    case SIGUSR2: caught_sig = CS_USR2; break;
-    default: internal_error( "caught signal not in set" );
-    }
-  }
-
-
-static void compress( const lzma_options & encoder_options, const int num_workers,
-                      int debug_level, int num_slots, int infd, int outfd,
-                      const Pretty_print & pp, const sigset_t *unblocked )
-  {
-  /*
-    We could wait for signals with either sigwait() or sigsuspend(). SUSv2
-    states about sigwait() that its effect on signal actions is unspecified.
-    SUSv3 still claims the same.
-
-    The SUSv2 description of sigsuspend() talks about both the thread and the
-    whole process being suspended until a signal arrives, although thread
-    suspension seems much more likely from the wording. They note that they
-    filed a clarification request for this. SUSv3 cleans this up and chooses
-    thread suspension which was more logical anyway.
-
-    I favor sigsuspend() because I need to re-raise SIGTERM and SIGINT, and
-    unspecified action behavior with sigwait() seems messy.
-
-    13-OCT-2009 lacos
-  */
-
-  if( verbosity >= 1 ) pp();
-
-  Muxer_arg muxer_arg;
-  muxer_arg.dictionary_size = encoder_options.dictionary_size;
-  muxer_arg.match_len_limit = encoder_options.match_len_limit;
-  muxer_arg.num_workers = num_workers;
-  muxer_arg.num_slots = num_slots;
-  muxer_arg.debug_level = debug_level;
-  muxer_arg.infd = infd;
-  muxer_arg.outfd = outfd;
-
-  pthread_t muxer_thread;
-  xcreate(&muxer_thread, muxer, &muxer_arg);
-
-  /* Unblock signals, wait for them, then block them again. */
-  {
-    int ret = sigsuspend(unblocked);
-    assert(-1 == ret && EINTR == errno);
-  }
-
-  switch( caught_sig ) {
-    case CS_INT:
-    case CS_TERM:	// FIXME remove output file
-      {
-        int sig;
-        sigset_t mask;
-
-        sig = (CS_INT == caught_sig) ? SIGINT : SIGTERM;
-        /*
-          We might have inherited a SIG_IGN from the parent, but that would
-          make no sense here. 24-OCT-2009 lacos
-        */
-        xsigaction(sig, SIG_DFL);
-        xraise(sig);
-
-        xsigemptyset(&mask);
-        xsigaddset(&mask, sig);
-        xsigmask(SIG_UNBLOCK, &mask, 0);
-      }
-      /*
-        We shouldn't reach this point, but if we do for some reason, fall
-        through.
-      */
-
-    case CS_USR1:
-      /* Error from a non-main thread via fatal(). */
-      fatal();
-
-    case CS_USR2:
-      /* Muxer thread joined other sub-threads and finished successfully. */
-      break;
-
-    default:
-      assert(0);
-  }
-
-  xjoin(muxer_thread);
-  }
-
-
-static void
-sigs_mod(int block_n_catch, sigset_t *oset)
-  {
-  void (*handler)(int);
-
-  if( block_n_catch ) {
-    sigset_t mask;
-
-    xsigemptyset(&mask);
-    xsigaddset(&mask, SIGINT);
-    xsigaddset(&mask, SIGTERM);
-    xsigaddset(&mask, SIGUSR1);
-    xsigaddset(&mask, SIGUSR2);
-    xsigmask(SIG_BLOCK, &mask, oset);
-
-    handler = sighandler;
-  }
-  else {
-    handler = SIG_DFL;
-  }
-
-  xsigaction(SIGINT,  handler);
-  xsigaction(SIGTERM, handler);
-  xsigaction(SIGUSR1, handler);
-  xsigaction(SIGUSR2, handler);
-
-  if( !block_n_catch ) {
-    xsigmask(SIG_SETMASK, oset, 0);
-  }
-  }
+// This can be called from any thread, main thread or sub-threads alike, since
+// they all call common helper functions that call fatal() in case of an error.
+//
+void fatal() { signal_handler( SIGUSR1 ); }
 
 
 // Returns the number of bytes really read.
 // If (returned value < size) and (errno == 0), means EOF was reached.
 //
-int readblock( const int fd, char * buf, const int size ) throw()
+int readblock( const int fd, uint8_t * buf, const int size ) throw()
   {
   int rest = size;
   errno = 0;
@@ -935,7 +589,7 @@ int readblock( const int fd, char * buf, const int size ) throw()
 // Returns the number of bytes really written.
 // If (returned value < size), it is always an error.
 //
-int writeblock( const int fd, const char * buf, const int size ) throw()
+int writeblock( const int fd, const uint8_t * buf, const int size ) throw()
   {
   int rest = size;
   errno = 0;
@@ -966,6 +620,7 @@ int main( const int argc, const char * argv[] )
     { 1 << 24, 163 },		// -8
     { 1 << 25, 273 } };		// -9
   lzma_options encoder_options = option_mapping[5];	// default = "-6"
+  int data_size = 0;
   int debug_level = 0;
   int inhandle = -1;
   int num_workers = 0;		// Start this many worker threads
@@ -977,22 +632,18 @@ int main( const int argc, const char * argv[] )
   std::string default_output_filename;
   std::vector< std::string > filenames;
   invocation_name = argv[0];
+  main_thread = pthread_self();
+  main_thread_pid = getpid();
 
   if( LZ_version()[0] != LZ_version_string[0] )
     internal_error( "bad library version" );
 
-  main_thread = pthread_self();
-  pid = getpid();
-
-  xsigaction(SIGPIPE, SIG_IGN);
-  xsigaction(SIGXFSZ, SIG_IGN);
-
   const int slots_per_worker = 2;
   long max_workers = sysconf( _SC_THREAD_THREADS_MAX );
   if( max_workers < 1 || max_workers > INT_MAX / slots_per_worker )
     max_workers = INT_MAX / slots_per_worker;
-  if( max_workers > INT_MAX / (int)sizeof( pthread_t ) )
-    max_workers = INT_MAX / sizeof( pthread_t );
+  if( max_workers > INT_MAX / (int)sizeof (pthread_t) )
+    max_workers = INT_MAX / sizeof (pthread_t);
 
   const Arg_parser::Option options[] =
     {
@@ -1006,6 +657,7 @@ int main( const int argc, const char * argv[] )
     { '8',  0,                Arg_parser::no  },
     { '9', "best",            Arg_parser::no  },
     { 'b', "member-size",     Arg_parser::yes },
+    { 'B', "data-size",       Arg_parser::yes },
     { 'c', "stdout",          Arg_parser::no  },
     { 'd', "decompress",      Arg_parser::no  },
     { 'D', "debug",           Arg_parser::yes },
@@ -1040,6 +692,8 @@ int main( const int argc, const char * argv[] )
       case '7': case '8': case '9':
                 encoder_options = option_mapping[code-'1']; break;
       case 'b': break;
+      case 'B': data_size = getnum( arg, 0, 100000,
+                                    2 * LZ_max_dictionary_size() ); break;
       case 'c': to_stdout = true; break;
       case 'd': program_mode = m_decompress; break;
       case 'D': debug_level = getnum( arg, 0, 0, 3 );
@@ -1048,8 +702,8 @@ int main( const int argc, const char * argv[] )
       case 'h': show_help(); return 0;
       case 'k': keep_input_files = true; break;
       case 'm': encoder_options.match_len_limit =
-                getnum( arg, 0, LZ_min_match_len_limit(),
-                                LZ_max_match_len_limit() ); break;
+                  getnum( arg, 0, LZ_min_match_len_limit(),
+                                  LZ_max_match_len_limit() ); break;
       case 'o': default_output_filename = arg; break;
       case 'n': num_workers = getnum( arg, 0, 1, max_workers ); break;
       case 'q': verbosity = -1; break;
@@ -1063,13 +717,16 @@ int main( const int argc, const char * argv[] )
       }
     }
 
+  if( data_size <= 0 )
+    data_size = 2 * std::max( 65536, encoder_options.dictionary_size );
+
   if( num_workers <= 0 )
     {
     long num_online = sysconf( _SC_NPROCESSORS_ONLN );
-    if( num_online <= 0 ) num_online = 2;
+    if( num_online <= 0 ) num_online = 1;
     num_workers = std::min( num_online, max_workers );
     }
-  const int num_slots = num_workers * slots_per_worker;
+  const int num_slots = std::max( 1, ( num_workers * slots_per_worker ) - 1 );
 
   bool filenames_given = false;
   for( ; argind < parser.arguments(); ++argind )
@@ -1079,7 +736,9 @@ int main( const int argc, const char * argv[] )
     }
 
   if( filenames.empty() ) filenames.push_back("-");
-  if( filenames_given && program_mode != m_compress ) set_signals();
+  if( !to_stdout && program_mode != m_test &&
+      ( filenames_given || default_output_filename.size() ) )
+    set_signals();
 
   Pretty_print pp( filenames );
   if( program_mode == m_test )
@@ -1144,14 +803,12 @@ int main( const int argc, const char * argv[] )
       delete_output_on_interrupt = true;
     const struct stat * const in_statsp = input_filename.size() ? &in_stats : 0;
     pp.set_name( input_filename );
+    if( verbosity >= 1 ) pp();
     int tmp = 0;
     if( program_mode == m_compress )
-      {
-      sigset_t unblocked;
-      sigs_mod(1, &unblocked);
-      compress( encoder_options, num_workers, debug_level, num_slots, inhandle, outhandle, pp, &unblocked );
-      sigs_mod(0, &unblocked);
-      }
+      tmp = compress( data_size, encoder_options.dictionary_size,
+                      encoder_options.match_len_limit, num_workers,
+                      num_slots, inhandle, outhandle, debug_level );
     else
       tmp = decompress( inhandle, pp, program_mode == m_test );
     if( tmp > retval ) retval = tmp;