X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=tests%2Fdissect-sysfile.c;h=444c15cc88bbec3544c017422aecb11b7b700bb3;hb=9ade26c8349b4434008c46cf09bc7473ec743972;hp=85f3644211d3171423239c59b14ab7136f28b36a;hpb=b5c82cc9aabe7e641011130240ae1b2e84348e23;p=pspp-builds.git

diff --git a/tests/dissect-sysfile.c b/tests/dissect-sysfile.c
index 85f36442..444c15cc 100644
--- a/tests/dissect-sysfile.c
+++ b/tests/dissect-sysfile.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc.
+   Copyright (C) 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -36,7 +36,7 @@
 #include "gettext.h"
 #define _(msgid) gettext (msgid)
 
-#define VAR_NAME_LEN 64
+#define ID_MAX_LEN 64
 
 struct sfm_reader
   {
@@ -45,8 +45,14 @@ struct sfm_reader
 
     int n_variable_records, n_variables;
 
+    int *var_widths;
+    size_t n_var_widths, allocated_var_widths;
+
     enum integer_format integer_format;
     enum float_format float_format;
+
+    bool compressed;
+    double bias;
   };
 
 static void read_header (struct sfm_reader *);
@@ -58,6 +64,7 @@ static void read_machine_integer_info (struct sfm_reader *,
                                        size_t size, size_t count);
 static void read_machine_float_info (struct sfm_reader *,
                                      size_t size, size_t count);
+static void read_mrsets (struct sfm_reader *, size_t size, size_t count);
 static void read_display_parameters (struct sfm_reader *,
                                      size_t size, size_t count);
 static void read_long_var_name_map (struct sfm_reader *r,
@@ -68,10 +75,14 @@ static void read_datafile_attributes (struct sfm_reader *r,
                                       size_t size, size_t count);
 static void read_variable_attributes (struct sfm_reader *r,
                                       size_t size, size_t count);
+static void read_ncases64 (struct sfm_reader *, size_t size, size_t count);
 static void read_character_encoding (struct sfm_reader *r,
 				       size_t size, size_t count);
 static void read_long_string_value_labels (struct sfm_reader *r,
                                            size_t size, size_t count);
+static void read_unknown_extension (struct sfm_reader *,
+                                    size_t size, size_t count);
+static void read_compressed_data (struct sfm_reader *);
 
 static struct text_record *open_text_record (
   struct sfm_reader *, size_t size);
@@ -80,6 +91,8 @@ static bool read_variable_to_value_pair (struct text_record *,
                                          char **key, char **value);
 static char *text_tokenize (struct text_record *, int delimiter);
 static bool text_match (struct text_record *text, int c);
+static const char *text_parse_counted_string (struct text_record *);
+static size_t text_pos (const struct text_record *);
 
 static void usage (int exit_code);
 static void sys_warn (struct sfm_reader *, const char *, ...)
@@ -89,7 +102,9 @@ static void sys_error (struct sfm_reader *, const char *, ...)
      NO_RETURN;
 
 static void read_bytes (struct sfm_reader *, void *, size_t);
+static bool try_read_bytes (struct sfm_reader *, void *, size_t);
 static int read_int (struct sfm_reader *);
+static int64_t read_int64 (struct sfm_reader *);
 static double read_float (struct sfm_reader *);
 static void read_string (struct sfm_reader *, char *, size_t);
 static void skip_bytes (struct sfm_reader *, size_t);
@@ -112,9 +127,13 @@ main (int argc, char *argv[])
       r.file_name = argv[i];
       r.file = fopen (r.file_name, "rb");
       if (r.file == NULL)
-        error (EXIT_FAILURE, errno, "error opening \"%s\"", r.file_name);
+        error (EXIT_FAILURE, errno, "error opening `%s'", r.file_name);
       r.n_variable_records = 0;
       r.n_variables = 0;
+      r.n_var_widths = 0;
+      r.allocated_var_widths = 0;
+      r.var_widths = 0;
+      r.compressed = false;
 
       if (argc > 2)
         printf ("Reading \"%s\":\n", r.file_name);
@@ -147,9 +166,13 @@ main (int argc, char *argv[])
               sys_error (&r, _("Unrecognized record type %d."), rec_type);
             }
         }
-      printf ("%08lx: end-of-dictionary record "
-              "(first byte of data at %08lx)\n",
-              ftell (r.file), ftell (r.file) + 4);
+      printf ("%08llx: end-of-dictionary record "
+              "(first byte of data at %08llx)\n",
+              (long long int) ftello (r.file),
+              (long long int) ftello (r.file) + 4);
+
+      if (r.compressed)
+        read_compressed_data (&r);
 
       fclose (r.file);
     }
@@ -169,7 +192,6 @@ read_header (struct sfm_reader *r)
   int32_t weight_index;
   int32_t ncases;
   uint8_t raw_bias[8];
-  double bias;
   char creation_date[10];
   char creation_time[9];
   char file_label[65];
@@ -193,10 +215,12 @@ read_header (struct sfm_reader *r)
                              raw_layout_code, sizeof raw_layout_code);
 
   nominal_case_size = read_int (r);
-  compressed = read_int (r) != 0;
+  compressed = read_int (r);
   weight_index = read_int (r);
   ncases = read_int (r);
 
+  r->compressed = compressed != 0;
+
   /* Identify floating-point format and obtain compression bias. */
   read_bytes (r, raw_bias, sizeof raw_bias);
   if (float_identify (100.0, raw_bias, sizeof raw_bias, &r->float_format) == 0)
@@ -209,7 +233,7 @@ read_header (struct sfm_reader *r)
       else
         r->float_format = FLOAT_IEEE_DOUBLE_LE;
     }
-  bias = float_get_double (r->float_format, raw_bias);
+  r->bias = float_get_double (r->float_format, raw_bias);
 
   read_string (r, creation_date, sizeof creation_date);
   read_string (r, creation_time, sizeof creation_time);
@@ -223,7 +247,7 @@ read_header (struct sfm_reader *r)
   printf ("\t%17s: %"PRId32"\n", "Compressed", compressed);
   printf ("\t%17s: %"PRId32"\n", "Weight index", weight_index);
   printf ("\t%17s: %"PRId32"\n", "Number of cases", ncases);
-  printf ("\t%17s: %g\n", "Compression bias", bias);
+  printf ("\t%17s: %g\n", "Compression bias", r->bias);
   printf ("\t%17s: %s\n", "Creation date", creation_date);
   printf ("\t%17s: %s\n", "Creation time", creation_time);
   printf ("\t%17s: \"%s\"\n", "File label", file_label);
@@ -287,8 +311,8 @@ read_variable_record (struct sfm_reader *r)
   int write_format;
   char name[9];
 
-  printf ("%08lx: variable record #%d\n",
-          ftell (r->file), r->n_variable_records++);
+  printf ("%08llx: variable record #%d\n",
+          (long long int) ftello (r->file), r->n_variable_records++);
 
   width = read_int (r);
   has_variable_label = read_int (r);
@@ -301,6 +325,11 @@ read_variable_record (struct sfm_reader *r)
   if (width >= 0)
     r->n_variables++;
 
+  if (r->n_var_widths >= r->allocated_var_widths)
+    r->var_widths = x2nrealloc (r->var_widths, &r->allocated_var_widths,
+                                sizeof *r->var_widths);
+  r->var_widths[r->n_var_widths++] = width;
+
   printf ("\tWidth: %d (%s)\n",
           width,
           width > 0 ? "string"
@@ -328,17 +357,21 @@ read_variable_record (struct sfm_reader *r)
     sys_error (r, _("Variable label indicator field is not 0 or 1."));
   if (has_variable_label == 1)
     {
-      long int offset = ftell (r->file);
-      size_t len;
+      long long int offset = ftello (r->file);
+      size_t len, read_len;
       char label[255 + 1];
 
       len = read_int (r);
-      if (len >= sizeof label)
-        sys_error (r, _("Variable %s has label of invalid length %zu."),
-                   name, len);
-      read_string (r, label, len + 1);
-      printf("\t%08lx Variable label: \"%s\"\n", offset, label);
 
+      /* Read up to 255 bytes of label. */
+      read_len = MIN (sizeof label - 1, len);
+      read_string (r, label, read_len + 1);
+      printf("\t%08llx Variable label: \"%s\"\n", offset, label);
+
+      /* Skip unread label bytes. */
+      skip_bytes (r, len - read_len);
+
+      /* Skip label padding up to multiple of 4 bytes. */
       skip_bytes (r, ROUND_UP (len, 4) - len);
     }
 
@@ -347,7 +380,7 @@ read_variable_record (struct sfm_reader *r)
     {
       int i;
 
-      printf ("\t%08lx Missing values:", ftell (r->file));
+      printf ("\t%08llx Missing values:", (long long int) ftello (r->file));
       if (!width)
         {
           if (missing_value_code < -3 || missing_value_code > 3
@@ -380,6 +413,20 @@ read_variable_record (struct sfm_reader *r)
     }
 }
 
+static void
+print_untyped_value (struct sfm_reader *r, char raw_value[8])
+{
+  int n_printable;
+  double value;
+
+  value = float_get_double (r->float_format, raw_value);
+  for (n_printable = 0; n_printable < sizeof raw_value; n_printable++)
+    if (!isprint (raw_value[n_printable]))
+      break;
+
+  printf ("%g/\"%.*s\"", value, n_printable, raw_value);
+}
+
 /* Reads value labels from sysfile R and inserts them into the
    associated dictionary. */
 static void
@@ -388,24 +435,18 @@ read_value_label_record (struct sfm_reader *r)
   int label_cnt, var_cnt;
   int i;
 
-  printf ("%08lx: value labels record\n", ftell (r->file));
+  printf ("%08llx: value labels record\n", (long long int) ftello (r->file));
 
   /* Read number of labels. */
   label_cnt = read_int (r);
   for (i = 0; i < label_cnt; i++)
     {
       char raw_value[8];
-      double value;
-      int n_printable;
       unsigned char label_len;
       size_t padded_len;
       char label[256];
 
       read_bytes (r, raw_value, sizeof raw_value);
-      value = float_get_double (r->float_format, raw_value);
-      for (n_printable = 0; n_printable < sizeof raw_value; n_printable++)
-        if (!isprint (raw_value[n_printable]))
-          break;
 
       /* Read label length. */
       read_bytes (r, &label_len, sizeof label_len);
@@ -415,7 +456,9 @@ read_value_label_record (struct sfm_reader *r)
       read_bytes (r, label, padded_len - 1);
       label[label_len] = 0;
 
-      printf ("\t%g/\"%.*s\": \"%s\"\n", value, n_printable, raw_value, label);
+      printf ("\t");
+      print_untyped_value (r, raw_value);
+      printf (": \"%s\"\n", label);
     }
 
   /* Now, read the type 4 record that has the list of variables
@@ -428,7 +471,7 @@ read_value_label_record (struct sfm_reader *r)
 
   /* Read number of variables associated with value label from type 4
      record. */
-  printf ("\t%08lx: apply to variables", ftell (r->file));
+  printf ("\t%08llx: apply to variables", (long long int) ftello (r->file));
   var_cnt = read_int (r);
   for (i = 0; i < var_cnt; i++)
     printf (" #%d", read_int (r));
@@ -441,14 +484,14 @@ read_document_record (struct sfm_reader *r)
   int n_lines;
   int i;
 
-  printf ("%08lx: document record\n", ftell (r->file));
+  printf ("%08llx: document record\n", (long long int) ftello (r->file));
   n_lines = read_int (r);
   printf ("\t%d lines of documents\n", n_lines);
 
   for (i = 0; i < n_lines; i++)
     {
       char line[81];
-      printf ("\t%08lx: ", ftell (r->file));
+      printf ("\t%08llx: ", (long long int) ftello (r->file));
       read_string (r, line, sizeof line);
       trim_spaces (line);
       printf ("line %d: \"%s\"\n", i, line);
@@ -458,13 +501,13 @@ read_document_record (struct sfm_reader *r)
 static void
 read_extension_record (struct sfm_reader *r)
 {
-  long int offset = ftell (r->file);
+  long long int offset = ftello (r->file);
   int subtype = read_int (r);
   size_t size = read_int (r);
   size_t count = read_int (r);
   size_t bytes = size * count;
 
-  printf ("%08lx: Record 7, subtype %d, size=%zu, count=%zu\n",
+  printf ("%08llx: Record 7, subtype %d, size=%zu, count=%zu\n",
           offset, subtype, size, count);
 
   switch (subtype)
@@ -489,8 +532,9 @@ read_extension_record (struct sfm_reader *r)
       break;
 
     case 7:
-      /* Unknown purpose. */
-      break;
+    case 19:
+      read_mrsets (r, size, count);
+      return;
 
     case 11:
       read_display_parameters (r, size, count);
@@ -505,8 +549,8 @@ read_extension_record (struct sfm_reader *r)
       return;
 
     case 16:
-      /* New in SPSS v14?  Unknown purpose.  */
-      break;
+      read_ncases64 (r, size, count);
+      return;
 
     case 17:
       read_datafile_attributes (r, size, count);
@@ -526,7 +570,8 @@ read_extension_record (struct sfm_reader *r)
 
     default:
       sys_warn (r, _("Unrecognized record type 7, subtype %d."), subtype);
-      break;
+      read_unknown_extension (r, size, count);
+      return;
     }
 
   skip_bytes (r, bytes);
@@ -535,7 +580,7 @@ read_extension_record (struct sfm_reader *r)
 static void
 read_machine_integer_info (struct sfm_reader *r, size_t size, size_t count)
 {
-  long int offset = ftell (r->file);
+  long long int offset = ftello (r->file);
   int version_major = read_int (r);
   int version_minor = read_int (r);
   int version_revision = read_int (r);
@@ -545,7 +590,7 @@ read_machine_integer_info (struct sfm_reader *r, size_t size, size_t count)
   int integer_representation = read_int (r);
   int character_code = read_int (r);
 
-  printf ("%08lx: machine integer info\n", offset);
+  printf ("%08llx: machine integer info\n", offset);
   if (size != 4 || count != 8)
     sys_error (r, _("Bad size (%zu) or count (%zu) field on record type 7, "
                     "subtype 3."),
@@ -571,12 +616,12 @@ read_machine_integer_info (struct sfm_reader *r, size_t size, size_t count)
 static void
 read_machine_float_info (struct sfm_reader *r, size_t size, size_t count)
 {
-  long int offset = ftell (r->file);
+  long long int offset = ftello (r->file);
   double sysmis = read_float (r);
   double highest = read_float (r);
   double lowest = read_float (r);
 
-  printf ("%08lx: machine float info\n", offset);
+  printf ("%08llx: machine float info\n", offset);
   if (size != 8 || count != 3)
     sys_error (r, _("Bad size (%zu) or count (%zu) on extension 4."),
                size, count);
@@ -597,6 +642,108 @@ read_machine_float_info (struct sfm_reader *r, size_t size, size_t count)
               lowest, "LOWEST");
 }
 
+/* Read record type 7, subtype 7. */
+static void
+read_mrsets (struct sfm_reader *r, size_t size, size_t count)
+{
+  struct text_record *text;
+
+  printf ("%08llx: multiple response sets\n",
+          (long long int) ftello (r->file));
+  text = open_text_record (r, size * count);
+  for (;;)
+    {
+      const char *name;
+      enum { MRSET_MC, MRSET_MD } type;
+      bool cat_label_from_counted_values = false;
+      bool label_from_var_label = false;
+      const char *counted;
+      const char *label;
+      const char *variables;
+
+      name = text_tokenize (text, '=');
+      if (name == NULL)
+        break;
+
+      if (text_match (text, 'C'))
+        {
+          type = MRSET_MC;
+          counted = NULL;
+          if (!text_match (text, ' '))
+            {
+              sys_warn (r, "missing space following 'C' at offset %zu "
+                        "in mrsets record", text_pos (text));
+              break;
+            }
+        }
+      else if (text_match (text, 'D'))
+        {
+          type = MRSET_MD;
+        }
+      else if (text_match (text, 'E'))
+        {
+          char *number;
+
+          type = MRSET_MD;
+          cat_label_from_counted_values = true;
+
+          if (!text_match (text, ' '))
+            {
+              sys_warn (r, _("Missing space following `%c' at offset %zu "
+                             "in MRSETS record"), 'E', text_pos (text));
+              break;
+            }
+
+          number = text_tokenize (text, ' ');
+          if (!strcmp (number, "11"))
+            label_from_var_label = true;
+          else if (strcmp (number, "1"))
+            sys_warn (r, _("Unexpected label source value `%s' "
+                           "following `E' at offset %zu in MRSETS record"),
+                      number, text_pos (text));
+
+        }
+      else
+        {
+          sys_warn (r, "missing `C', `D', or `E' at offset %zu "
+                    "in mrsets record", text_pos (text));
+          break;
+        }
+
+      if (type == MRSET_MD)
+        {
+          counted = text_parse_counted_string (text);
+          if (counted == NULL)
+            break;
+        }
+
+      label = text_parse_counted_string (text);
+      if (label == NULL)
+        break;
+
+      variables = text_tokenize (text, '\n');
+      if (variables == NULL)
+        {
+          sys_warn (r, "missing variable names following label "
+                    "at offset %zu in mrsets record", text_pos (text));
+          break;
+        }
+
+      printf ("\t\"%s\": multiple %s set",
+              name, type == MRSET_MC ? "category" : "dichotomy");
+      if (counted != NULL)
+        printf (", counted value \"%s\"", counted);
+      if (cat_label_from_counted_values)
+        printf (", category labels from counted values");
+      if (label[0] != '\0')
+        printf (", label \"%s\"", label);
+      if (label_from_var_label)
+        printf (", label from variable label");
+      printf(", variables \"%s\"\n", variables);
+    }
+  close_text_record (text);
+}
+
 /* Read record type 7, subtype 11. */
 static void
 read_display_parameters (struct sfm_reader *r, size_t size, size_t count)
@@ -605,7 +752,8 @@ read_display_parameters (struct sfm_reader *r, size_t size, size_t count)
   bool includes_width;
   size_t i;
 
-  printf ("%08lx: variable display parameters\n", ftell (r->file));
+  printf ("%08llx: variable display parameters\n",
+          (long long int) ftello (r->file));
   if (size != 4)
     {
       sys_warn (r, _("Bad size %zu on extension 11."), size);
@@ -656,7 +804,8 @@ read_long_var_name_map (struct sfm_reader *r, size_t size, size_t count)
   char *var;
   char *long_name;
 
-  printf ("%08lx: long variable names (short => long)\n", ftell (r->file));
+  printf ("%08llx: long variable names (short => long)\n",
+          (long long int) ftello (r->file));
   text = open_text_record (r, size * count);
   while (read_variable_to_value_pair (text, &var, &long_name))
     printf ("\t%s => %s\n", var, long_name);
@@ -672,7 +821,8 @@ read_long_string_map (struct sfm_reader *r, size_t size, size_t count)
   char *var;
   char *length_s;
 
-  printf ("%08lx: very long strings (variable => length)\n", ftell (r->file));
+  printf ("%08llx: very long strings (variable => length)\n",
+          (long long int) ftello (r->file));
   text = open_text_record (r, size * count);
   while (read_variable_to_value_pair (text, &var, &length_s))
     printf ("\t%s => %d\n", var, atoi (length_s));
@@ -720,12 +870,37 @@ read_attributes (struct sfm_reader *r, struct text_record *text,
     }
 }
 
+/* Read extended number of cases record. */
+static void
+read_ncases64 (struct sfm_reader *r, size_t size, size_t count)
+{
+  int64_t unknown, ncases64;
+
+  if (size != 8)
+    {
+      sys_warn (r, _("Bad size %zu for extended number of cases."), size);
+      skip_bytes (r, size * count);
+      return;
+    }
+  if (count != 2)
+    {
+      sys_warn (r, _("Bad count %zu for extended number of cases."), size);
+      skip_bytes (r, size * count);
+      return;
+    }
+  unknown = read_int64 (r);
+  ncases64 = read_int64 (r);
+  printf ("%08llx: extended number of cases: "
+          "unknown=%"PRId64", ncases64=%"PRId64"\n",
+          (long long int) ftello (r->file), unknown, ncases64);
+}
+
 static void
 read_datafile_attributes (struct sfm_reader *r, size_t size, size_t count) 
 {
   struct text_record *text;
   
-  printf ("%08lx: datafile attributes\n", ftell (r->file));
+  printf ("%08llx: datafile attributes\n", (long long int) ftello (r->file));
   text = open_text_record (r, size * count);
   read_attributes (r, text, "datafile");
   close_text_record (text);
@@ -734,23 +909,23 @@ read_datafile_attributes (struct sfm_reader *r, size_t size, size_t count)
 static void
 read_character_encoding (struct sfm_reader *r, size_t size, size_t count)
 {
-  const unsigned long int posn =  ftell (r->file);
+  long long int posn =  ftello (r->file);
   char *encoding = xcalloc (size, count + 1);
   read_string (r, encoding, count + 1);
 
-  printf ("%08lx: Character Encoding: %s\n", posn, encoding);
+  printf ("%08llx: Character Encoding: %s\n", posn, encoding);
 }
 
 static void
 read_long_string_value_labels (struct sfm_reader *r, size_t size, size_t count)
 {
-  const long start = ftell (r->file);
+  long long int start = ftello (r->file);
 
-  printf ("%08lx: long string value labels\n", start);
-  while (ftell (r->file) - start < size * count)
+  printf ("%08llx: long string value labels\n", start);
+  while (ftello (r->file) - start < size * count)
     {
-      long posn = ftell (r->file);
-      char var_name[VAR_NAME_LEN + 1];
+      long long posn = ftello (r->file);
+      char var_name[ID_MAX_LEN + 1];
       int var_name_len;
       int n_values;
       int width;
@@ -758,17 +933,17 @@ read_long_string_value_labels (struct sfm_reader *r, size_t size, size_t count)
 
       /* Read variable name. */
       var_name_len = read_int (r);
-      if (var_name_len > VAR_NAME_LEN)
+      if (var_name_len > ID_MAX_LEN)
         sys_error (r, _("Variable name length in long string value label "
                         "record (%d) exceeds %d-byte limit."),
-                   var_name_len, VAR_NAME_LEN);
+                   var_name_len, ID_MAX_LEN);
       read_string (r, var_name, var_name_len + 1);
 
       /* Read width, number of values. */
       width = read_int (r);
       n_values = read_int (r);
 
-      printf ("\t%08lx: %s, width %d, %d values\n",
+      printf ("\t%08llx: %s, width %d, %d values\n",
               posn, var_name, width, n_values);
 
       /* Read values. */
@@ -780,7 +955,7 @@ read_long_string_value_labels (struct sfm_reader *r, size_t size, size_t count)
           char *label;
 	  int label_length;
 
-          posn = ftell (r->file);
+          posn = ftello (r->file);
 
           /* Read value. */
           value_length = read_int (r);
@@ -792,7 +967,7 @@ read_long_string_value_labels (struct sfm_reader *r, size_t size, size_t count)
           label = xmalloc (label_length + 1);
           read_string (r, label, label_length + 1);
 
-          printf ("\t\t%08lx: \"%s\" (%d bytes) => \"%s\" (%d bytes)\n",
+          printf ("\t\t%08llx: \"%s\" (%d bytes) => \"%s\" (%d bytes)\n",
                   posn, value, value_length, label, label_length);
 
           free (value);
@@ -801,12 +976,86 @@ read_long_string_value_labels (struct sfm_reader *r, size_t size, size_t count)
     }
 }
 
+static void
+hex_dump (size_t offset, const void *buffer_, size_t buffer_size)
+{
+  const uint8_t *buffer = buffer_;
+
+  while (buffer_size > 0)
+    {
+      size_t n = MIN (buffer_size, 16);
+      size_t i;
+
+      printf ("%04zx", offset);
+      for (i = 0; i < 16; i++)
+        {
+          if (i < n)
+            printf ("%c%02x", i == 8 ? '-' : ' ', buffer[i]);
+          else
+            printf ("   ");
+        }
+
+      printf (" |");
+      for (i = 0; i < 16; i++)
+        {
+          unsigned char c = i < n ? buffer[i] : ' ';
+          putchar (isprint (c) ? c : '.');
+        }
+      printf ("|\n");
+
+      offset += n;
+      buffer += n;
+      buffer_size -= n;
+    }
+}
+
+/* Reads and prints any type 7 record that we don't understand. */
+static void
+read_unknown_extension (struct sfm_reader *r, size_t size, size_t count)
+{
+  unsigned char *buffer;
+  size_t i;
+
+  if (size == 0 || count > 65536 / size)
+    skip_bytes (r, size * count);
+  else if (size != 1)
+    {
+      buffer = xmalloc (size);
+      for (i = 0; i < count; i++)
+        {
+          read_bytes (r, buffer, size);
+          hex_dump (i * size, buffer, size);
+        }
+      free (buffer);
+    }
+  else
+    {
+      buffer = xmalloc (count);
+      read_bytes (r, buffer, count);
+      if (memchr (buffer, 0, count) == 0)
+        for (i = 0; i < count; i++)
+          {
+            unsigned char c = buffer[i];
+
+            if (c == '\\')
+              printf ("\\\\");
+            else if (c == '\n' || isprint (c))
+              putchar (c);
+            else
+              printf ("\\%02x", c);
+          }
+      else
+        hex_dump (0, buffer, count);
+      free (buffer);
+    }
+}
+
 static void
 read_variable_attributes (struct sfm_reader *r, size_t size, size_t count) 
 {
   struct text_record *text;
   
-  printf ("%08lx: variable attributes\n", ftell (r->file));
+  printf ("%08llx: variable attributes\n", (long long int) ftello (r->file));
   text = open_text_record (r, size * count);
   for (;;) 
     {
@@ -816,6 +1065,96 @@ read_variable_attributes (struct sfm_reader *r, size_t size, size_t count)
     }
   close_text_record (text);
 }
+
+static void
+read_compressed_data (struct sfm_reader *r)
+{
+  enum { N_OPCODES = 8 };
+  uint8_t opcodes[N_OPCODES];
+  long long int opcode_ofs;
+  int opcode_idx;
+  int case_num;
+  int i;
+
+  read_int (r);
+  printf ("\n%08llx: compressed data:\n", (long long int) ftello (r->file));
+
+  opcode_idx = N_OPCODES;
+  opcode_ofs = 0;
+  case_num = 0;
+  for (case_num = 0; ; case_num++)
+    {
+      printf ("%08llx: case %d's uncompressible data begins\n",
+              (long long int) ftello (r->file), case_num);
+      for (i = 0; i < r->n_var_widths; )
+        {
+          int width = r->var_widths[i];
+          char raw_value[8];
+          int opcode;
+
+          if (opcode_idx >= N_OPCODES)
+            {
+              opcode_ofs = ftello (r->file);
+              if (i == 0)
+                {
+                  if (!try_read_bytes (r, opcodes, 8))
+                    return;
+                }
+              else
+                read_bytes (r, opcodes, 8);
+              opcode_idx = 0;
+            }
+          opcode = opcodes[opcode_idx];
+          printf ("%08llx: variable %d: opcode %d: ",
+                  opcode_ofs + opcode_idx, i, opcode);
+
+          switch (opcode)
+            {
+            default:
+              printf ("%g", opcode - r->bias);
+              if (width != 0)
+                printf (", but this is a string variable (width=%d)", width);
+              printf ("\n");
+              i++;
+              break;
+
+            case 0:
+              printf ("ignored padding\n");
+              break;
+
+            case 252:
+              printf ("end of data\n");
+              return;
+
+            case 253:
+              read_bytes (r, raw_value, 8);
+              printf ("uncompressible data: ");
+              print_untyped_value (r, raw_value);
+              printf ("\n");
+              i++;
+              break;
+
+            case 254:
+              printf ("spaces");
+              if (width == 0)
+                printf (", but this is a numeric variable");
+              printf ("\n");
+              i++;
+              break;
+
+            case 255:
+              printf ("SYSMIS");
+              if (width != 0)
+                printf (", but this is a string variable (width=%d)", width);
+              printf ("\n");
+              i++;
+              break;
+            }
+
+          opcode_idx++;
+        }
+    }
+}
 
 /* Helpers for reading records that consist of structured text
    strings. */
@@ -823,6 +1162,7 @@ read_variable_attributes (struct sfm_reader *r, size_t size, size_t count)
 /* State. */
 struct text_record
   {
+    struct sfm_reader *reader;  /* Reader. */
     char *buffer;               /* Record contents. */
     size_t size;                /* Size of buffer. */
     size_t pos;                 /* Current position in buffer. */
@@ -836,6 +1176,8 @@ open_text_record (struct sfm_reader *r, size_t size)
   struct text_record *text = xmalloc (sizeof *text);
   char *buffer = xmalloc (size + 1);
   read_bytes (r, buffer, size);
+  buffer[size] = '\0';
+  text->reader = r;
   text->buffer = buffer;
   text->size = size;
   text->pos = 0;
@@ -878,6 +1220,54 @@ text_match (struct text_record *text, int c)
     return false;
 }
 
+/* Reads a integer value expressed in decimal, then a space, then a string that
+   consists of exactly as many bytes as specified by the integer, then a space,
+   from TEXT.  Returns the string, null-terminated, as a subset of TEXT's
+   buffer (so the caller should not free the string). */
+static const char *
+text_parse_counted_string (struct text_record *text)
+{
+  size_t start;
+  size_t n;
+  char *s;
+
+  start = text->pos;
+  n = 0;
+  while (isdigit ((unsigned char) text->buffer[text->pos]))
+    n = (n * 10) + (text->buffer[text->pos++] - '0');
+  if (start == text->pos)
+    {
+      sys_error (text->reader, "expecting digit at offset %zu in record",
+                 text->pos);
+      return NULL;
+    }
+
+  if (!text_match (text, ' '))
+    {
+      sys_error (text->reader, "expecting space at offset %zu in record",
+                 text->pos);
+      return NULL;
+    }
+
+  if (text->pos + n > text->size)
+    {
+      sys_error (text->reader, "%zu-byte string starting at offset %zu "
+                 "exceeds record length %zu", n, text->pos, text->size);
+      return NULL;
+    }
+
+  s = &text->buffer[text->pos];
+  if (s[n] != ' ')
+    {
+      sys_error (text->reader, "expecting space at offset %zu following "
+                 "%zu-byte string", text->pos + n, n);
+      return NULL;
+    }
+  s[n] = '\0';
+  text->pos += n + 1;
+  return s;
+}
+
 /* Reads a variable=value pair from TEXT.
    Looks up the variable in DICT and stores it into *VAR.
    Stores a null-terminated value into *VALUE. */
@@ -896,6 +1286,13 @@ read_variable_to_value_pair (struct text_record *text,
     text->pos++;
   return true;
 }
+
+/* Returns the current byte offset inside the TEXT's string. */
+static size_t
+text_pos (const struct text_record *text)
+{
+  return text->pos;
+}
 
 static void
 usage (int exit_code)
@@ -910,8 +1307,8 @@ usage (int exit_code)
 static void
 sys_msg (struct sfm_reader *r, const char *format, va_list args)
 {
-  printf ("\"%s\" near offset 0x%lx: ",
-          r->file_name, (unsigned long) ftell (r->file));
+  printf ("\"%s\" near offset 0x%llx: ",
+          r->file_name, (long long int) ftello (r->file));
   vprintf (format, args);
   putchar ('\n');
 }
@@ -971,6 +1368,16 @@ read_bytes (struct sfm_reader *r, void *buf, size_t byte_cnt)
   read_bytes_internal (r, false, buf, byte_cnt);
 }
 
+/* Reads BYTE_CNT bytes into BUF.
+   Returns true if exactly BYTE_CNT bytes are successfully read.
+   Returns false if an immediate end-of-file is encountered.
+   Aborts if an I/O error or a partial read occurs. */
+static bool
+try_read_bytes (struct sfm_reader *r, void *buf, size_t byte_cnt)
+{
+  return read_bytes_internal (r, true, buf, byte_cnt);
+}
+
 /* Reads a 32-bit signed integer from R and returns its value in
    host format. */
 static int
@@ -981,6 +1388,16 @@ read_int (struct sfm_reader *r)
   return integer_get (r->integer_format, integer, sizeof integer);
 }
 
+/* Reads a 64-bit signed integer from R and returns its value in
+   host format. */
+static int64_t
+read_int64 (struct sfm_reader *r)
+{
+  uint8_t integer[8];
+  read_bytes (r, integer, sizeof integer);
+  return integer_get (r->integer_format, integer, sizeof integer);
+}
+
 /* Reads a 64-bit floating-point number from R and returns its
    value in host format. */
 static double