X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fcasefile.c;h=8fe0740371aee2eab107119a46789e99ede08a64;hb=53ceff2865473a6b561b521986fafd31a993a1a6;hp=eca07f83e2610fcdb499372e661209dbb764e67b;hpb=cb962ee9edc95f73507d35f0714ec8aa68c5295c;p=pspp

diff --git a/src/casefile.c b/src/casefile.c
index eca07f83e2..8fe0740371 100644
--- a/src/casefile.c
+++ b/src/casefile.c
@@ -14,8 +14,8 @@
 
    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
-   Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
-   02111-1307, USA. */
+   Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
+   02110-1301, USA. */
 
 #include <config.h>
 #include "casefile.h"
@@ -29,22 +29,87 @@
 #include "alloc.h"
 #include "case.h"
 #include "error.h"
+#include "full-read.h"
+#include "full-write.h"
 #include "misc.h"
+#include "mkfile.h"
 #include "settings.h"
 #include "var.h"
 
-#ifdef HAVE_VALGRIND_VALGRIND_H
-#include <valgrind/valgrind.h>
-#endif
+#include "gettext.h"
+#define _(msgid) gettext (msgid)
 
 #define IO_BUF_SIZE (8192 / sizeof (union value))
 
-/* A casefile is a sequentially accessible array of immutable
-   cases.  It may be stored in memory or on disk as workspace
-   allows.  Cases may be appended to the end of the file.  Cases
-   may be read sequentially starting from the beginning of the
-   file.  Once any cases have been read, no more cases may be
-   appended.  The entire file is discarded at once. */
+/* A casefile represents a sequentially accessible stream of
+   immutable cases.
+
+   If workspace allows, a casefile is maintained in memory.  If
+   workspace overflows, then the casefile is pushed to disk.  In
+   either case the interface presented to callers is kept the
+   same.
+
+   The life cycle of a casefile consists of up to three phases:
+
+       1. Writing.  The casefile initially contains no cases.  In
+          this phase, any number of cases may be appended to the
+          end of a casefile.  (Cases are never inserted in the
+          middle or before the beginning of a casefile.)
+
+          Use casefile_append() or casefile_append_xfer() to
+          append a case to a casefile.
+
+       2. Reading.  The casefile may be read sequentially,
+          starting from the beginning, by "casereaders".  Any
+          number of casereaders may be created, at any time,
+          during the reading phase.  Each casereader has an
+          independent position in the casefile.
+
+          Casereaders may only move forward.  They cannot move
+          backward to arbitrary records or seek randomly.
+          Cloning casereaders is possible, but it is not yet
+          implemented.
+
+          Use casefile_get_reader() to create a casereader for
+          use in phase 2.  This also transitions from phase 1 to
+          phase 2.  Calling casefile_mode_reader() makes the same
+          transition, without creating a casereader.
+
+          Use casereader_read(), casereader_read_xfer(), or
+          casereader_read_xfer_assert() to read a case from a
+          casereader.  Use casereader_destroy() to discard a
+          casereader when it is no longer needed.
+
+       3. Destruction.  This phase is optional.  The casefile is
+          also read with casereaders in this phase, but the
+          ability to create new casereaders is curtailed.
+
+          In this phase, casereaders could still be cloned (once
+          we eventually implement cloning).
+
+          To transition from phase 1 or 2 to phase 3 and create a
+          casereader, call casefile_get_destructive_reader().
+          The same functions apply to the casereader obtained
+          this way as apply to casereaders obtained in phase 2.
+          
+          After casefile_get_destructive_reader() is called, no
+          more casereaders may be created with
+          casefile_get_reader() or
+          casefile_get_destructive_reader().  (If cloning of
+          casereaders were implemented, it would still be
+          possible.)
+
+          The purpose of the limitations applied to casereaders
+          in phase 3 is to allow in-memory casefiles to fully
+          transfer ownership of cases to the casereaders,
+          avoiding the need for extra copies of case data.  For
+          relatively static data sets with many variables, I
+          suspect (without evidence) that this may be a big
+          performance boost.
+
+   When a casefile is no longer needed, it may be destroyed with
+   casefile_destroy().  This function will also destroy any
+   remaining casereaders. */
 
 /* In-memory cases are arranged in an array of arrays.  The top
    level is variable size and the size of each bottom level array
@@ -113,8 +178,6 @@ static void fill_buffer (struct casereader *reader);
 
 static int safe_open (const char *filename, int flags);
 static int safe_close (int fd);
-static int full_read (int fd, void *buffer, size_t size);
-static int full_write (int fd, const void *buffer, size_t size);
 
 /* Creates and returns a casefile to store cases of VALUE_CNT
    `union value's each. */
@@ -268,7 +331,7 @@ casefile_append (struct casefile *cf, const struct ccase *c)
   /* Try memory first. */
   if (cf->storage == MEMORY) 
     {
-      if (case_bytes < get_max_workspace ())
+      if (case_bytes < get_workspace ())
         {
           size_t block_idx = cf->case_cnt / CASES_PER_BLOCK;
           size_t case_idx = cf->case_cnt % CASES_PER_BLOCK;
@@ -281,12 +344,12 @@ casefile_append (struct casefile *cf, const struct ccase *c)
               if ((block_idx & (block_idx - 1)) == 0) 
                 {
                   size_t block_cap = block_idx == 0 ? 1 : block_idx * 2;
-                  cf->cases = xrealloc (cf->cases,
-                                        sizeof *cf->cases * block_cap);
+                  cf->cases = xnrealloc (cf->cases,
+                                         block_cap, sizeof *cf->cases);
                 }
 
-              cf->cases[block_idx] = xmalloc (sizeof **cf->cases
-                                              * CASES_PER_BLOCK);
+              cf->cases[block_idx] = xnmalloc (CASES_PER_BLOCK,
+                                               sizeof **cf->cases);
             }
 
           case_move (&cf->cases[block_idx][case_idx], &new_case);
@@ -339,35 +402,6 @@ flush_buffer (struct casefile *cf)
     } 
 }
 
-/* Creates a temporary file and stores its name in *FILENAME and
-   a file descriptor for it in *FD.  Returns success.  Caller is
-   responsible for freeing *FILENAME. */
-static int
-make_temp_file (int *fd, char **filename)
-{
-  const char *parent_dir;
-
-  assert (filename != NULL);
-  assert (fd != NULL);
-
-  if (getenv ("TMPDIR") != NULL)
-    parent_dir = getenv ("TMPDIR");
-  else
-    parent_dir = P_tmpdir;
-
-  *filename = xmalloc (strlen (parent_dir) + 32);
-  sprintf (*filename, "%s%cpsppXXXXXX", parent_dir, DIR_SEPARATOR);
-  *fd = mkstemp (*filename);
-  if (*fd < 0)
-    {
-      msg (FE, _("%s: Creating temporary file: %s."),
-           *filename, strerror (errno));
-      free (*filename);
-      *filename = NULL;
-      return 0;
-    }
-  return 1;
-}
 
 /* If CF is currently stored in memory, writes it to disk.  Readers, if any,
    retain their current positions. */
@@ -390,7 +424,7 @@ casefile_to_disk (const struct casefile *cf_)
       cf->storage = DISK;
       if (!make_temp_file (&cf->fd, &cf->filename))
         err_failure ();
-      cf->buffer = xmalloc (cf->buffer_size * sizeof *cf->buffer);
+      cf->buffer = xnmalloc (cf->buffer_size, sizeof *cf->buffer);
       memset (cf->buffer, 0, cf->buffer_size * sizeof *cf->buffer);
 
       case_bytes -= cf->case_cnt * cf->case_acct_size;
@@ -515,7 +549,7 @@ reader_open_file (struct casereader *reader)
     }
   else 
     {
-      reader->buffer = xmalloc (cf->buffer_size * sizeof *cf->buffer);
+      reader->buffer = xnmalloc (cf->buffer_size, sizeof *cf->buffer);
       memset (reader->buffer, 0, cf->buffer_size * sizeof *cf->buffer); 
     }
 
@@ -563,7 +597,8 @@ casereader_get_casefile (const struct casereader *reader)
 }
 
 /* Reads a copy of the next case from READER into C.
-   Caller is responsible for destroying C. */
+   Caller is responsible for destroying C.
+   Returns true if successful, false at end of file. */
 int
 casereader_read (struct casereader *reader, struct ccase *c) 
 {
@@ -600,7 +635,8 @@ casereader_read (struct casereader *reader, struct ccase *c)
 }
 
 /* Reads the next case from READER into C and transfers ownership
-   to the caller.  Caller is responsible for destroying C. */
+   to the caller.  Caller is responsible for destroying C.
+   Returns true if successful, false at end of file. */
 int
 casereader_read_xfer (struct casereader *reader, struct ccase *c)
 {
@@ -622,6 +658,16 @@ casereader_read_xfer (struct casereader *reader, struct ccase *c)
     }
 }
 
+/* Reads the next case from READER into C and transfers ownership
+   to the caller.  Caller is responsible for destroying C.
+   Assert-fails at end of file. */
+void
+casereader_read_xfer_assert (struct casereader *reader, struct ccase *c) 
+{
+  bool success = casereader_read_xfer (reader, c);
+  assert (success);
+}
+
 /* Destroys READER. */
 void
 casereader_destroy (struct casereader *reader)
@@ -684,48 +730,6 @@ static int safe_close (int fd)
   return retval;
 }
 
-/* Calls read(), passing FD, BUFFER, and SIZE, repeating as
-   necessary to deal with interrupted calls. */
-static int
-full_read (int fd, void *buffer_, size_t size) 
-{
-  char *buffer = buffer_;
-  size_t bytes_read = 0;
-  
-  while (bytes_read < size)
-    {
-      int retval = read (fd, buffer + bytes_read, size - bytes_read);
-      if (retval > 0) 
-        bytes_read += retval; 
-      else if (retval == 0) 
-        return bytes_read;
-      else if (errno != EINTR)
-        return -1;
-    }
-
-  return bytes_read;
-}
-
-/* Calls write(), passing FD, BUFFER, and SIZE, repeating as
-   necessary to deal with interrupted calls. */
-static int
-full_write (int fd, const void *buffer_, size_t size) 
-{
-  const char *buffer = buffer_;
-  size_t bytes_written = 0;
-  
-  while (bytes_written < size)
-    {
-      int retval = write (fd, buffer + bytes_written, size - bytes_written);
-      if (retval >= 0) 
-        bytes_written += retval; 
-      else if (errno != EINTR)
-        return -1;
-    }
-
-  return bytes_written;
-}
-
 /* Registers our exit handler with atexit() if it has not already
    been registered. */
 static void
@@ -739,6 +743,8 @@ register_atexit (void)
     }
 }
 
+
+
 /* atexit() handler that closes and deletes our temporary
    files. */
 static void
@@ -747,194 +753,3 @@ exit_handler (void)
   while (casefiles != NULL)
     casefile_destroy (casefiles);
 }
-
-#include <gsl/gsl_rng.h>
-#include <stdarg.h>
-#include "command.h"
-#include "lexer.h"
-
-static void test_casefile (int pattern, size_t value_cnt, size_t case_cnt);
-static void get_random_case (struct ccase *, size_t value_cnt,
-                             size_t case_idx);
-static void write_random_case (struct casefile *cf, size_t case_idx);
-static void read_and_verify_random_case (struct casefile *cf,
-                                         struct casereader *reader,
-                                         size_t case_idx);
-static void fail_test (const char *message, ...);
-
-int
-cmd_debug_casefile (void) 
-{
-  static const size_t sizes[] =
-    {
-      1, 2, 3, 4, 5, 6, 7, 14, 15, 16, 17, 31, 55, 73,
-      100, 137, 257, 521, 1031, 2053
-    };
-  int size_max;
-  int case_max;
-  int pattern;
-
-  size_max = sizeof sizes / sizeof *sizes;
-  if (lex_match_id ("SMALL")) 
-    {
-      size_max -= 4;
-      case_max = 511; 
-    }
-  else
-    case_max = 4095;
-  if (token != '.')
-    return lex_end_of_command ();
-    
-  for (pattern = 0; pattern < 6; pattern++) 
-    {
-      const size_t *size;
-
-      for (size = sizes; size < sizes + size_max; size++) 
-        {
-          size_t case_cnt;
-
-          for (case_cnt = 0; case_cnt <= case_max;
-               case_cnt = (case_cnt * 2) + 1)
-            test_casefile (pattern, *size, case_cnt);
-        }
-    }
-  printf ("Casefile tests succeeded.\n");
-  return CMD_SUCCESS;
-}
-
-static void
-test_casefile (int pattern, size_t value_cnt, size_t case_cnt) 
-{
-  struct casefile *cf;
-  struct casereader *r1, *r2;
-  struct ccase c;
-  gsl_rng *rng;
-  size_t i, j;
-
-  rng = gsl_rng_alloc (gsl_rng_mt19937);
-  cf = casefile_create (value_cnt);
-  if (pattern == 5)
-    casefile_to_disk (cf);
-  for (i = 0; i < case_cnt; i++)
-    write_random_case (cf, i);
-  if (pattern == 5)
-    casefile_sleep (cf);
-  r1 = casefile_get_reader (cf);
-  r2 = casefile_get_reader (cf);
-  switch (pattern) 
-    {
-    case 0:
-    case 5:
-      for (i = 0; i < case_cnt; i++) 
-        {
-          read_and_verify_random_case (cf, r1, i);
-          read_and_verify_random_case (cf, r2, i);
-        } 
-      break;
-    case 1:
-      for (i = 0; i < case_cnt; i++)
-        read_and_verify_random_case (cf, r1, i);
-      for (i = 0; i < case_cnt; i++) 
-        read_and_verify_random_case (cf, r2, i);
-      break;
-    case 2:
-    case 3:
-    case 4:
-      for (i = j = 0; i < case_cnt; i++) 
-        {
-          read_and_verify_random_case (cf, r1, i);
-          if (gsl_rng_get (rng) % pattern == 0) 
-            read_and_verify_random_case (cf, r2, j++); 
-          if (i == case_cnt / 2)
-            casefile_to_disk (cf);
-        }
-      for (; j < case_cnt; j++) 
-        read_and_verify_random_case (cf, r2, j);
-      break;
-    }
-  if (casereader_read (r1, &c))
-    fail_test ("Casereader 1 not at end of file.");
-  if (casereader_read (r2, &c))
-    fail_test ("Casereader 2 not at end of file.");
-  if (pattern != 1)
-    casereader_destroy (r1);
-  if (pattern != 2)
-    casereader_destroy (r2);
-  if (pattern > 2) 
-    {
-      r1 = casefile_get_destructive_reader (cf);
-      for (i = 0; i < case_cnt; i++) 
-        {
-          struct ccase read_case, expected_case;
-          
-          get_random_case (&expected_case, value_cnt, i);
-          if (!casereader_read_xfer (r1, &read_case)) 
-            fail_test ("Premature end of casefile.");
-          for (j = 0; j < value_cnt; j++) 
-            {
-              double a = case_num (&read_case, j);
-              double b = case_num (&expected_case, j);
-              if (a != b)
-                fail_test ("Case %lu fails comparison.", (unsigned long) i); 
-            }
-          case_destroy (&expected_case);
-          case_destroy (&read_case);
-        }
-      casereader_destroy (r1);
-    }
-  casefile_destroy (cf);
-  gsl_rng_free (rng);
-}
-
-static void
-get_random_case (struct ccase *c, size_t value_cnt, size_t case_idx) 
-{
-  int i;
-  case_create (c, value_cnt);
-  for (i = 0; i < value_cnt; i++)
-    case_data_rw (c, i)->f = case_idx % 257 + i;
-}
-
-static void
-write_random_case (struct casefile *cf, size_t case_idx) 
-{
-  struct ccase c;
-  get_random_case (&c, casefile_get_value_cnt (cf), case_idx);
-  casefile_append_xfer (cf, &c);
-}
-
-static void
-read_and_verify_random_case (struct casefile *cf,
-                             struct casereader *reader, size_t case_idx) 
-{
-  struct ccase read_case, expected_case;
-  size_t value_cnt;
-  size_t i;
-  
-  value_cnt = casefile_get_value_cnt (cf);
-  get_random_case (&expected_case, value_cnt, case_idx);
-  if (!casereader_read (reader, &read_case)) 
-    fail_test ("Premature end of casefile.");
-  for (i = 0; i < value_cnt; i++) 
-    {
-      double a = case_num (&read_case, i);
-      double b = case_num (&expected_case, i);
-      if (a != b)
-        fail_test ("Case %lu fails comparison.", (unsigned long) case_idx); 
-    }
-  case_destroy (&read_case);
-  case_destroy (&expected_case);
-}
-
-static void
-fail_test (const char *message, ...) 
-{
-  va_list args;
-
-  va_start (args, message);
-  vprintf (message, args);
-  putchar ('\n');
-  va_end (args);
-  
-  exit (1);
-}