Figure out more parts of format.

[pspp] / dump.c

diff --git a/dump.c b/dump.c
index 6cc9e52abe4e5a9767b5d008841ba462a14893cc..12ee73ebdf4d02d143776cb252eca089ac215a7e 100644 (file)
--- a/dump.c
+++ b/dump.c
@@ -6,6 +6,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <sys/stat.h>
+#include <time.h>
 #include <unistd.h>
 #include "u8-mbtouc.h"
 
@@ -63,6 +64,15 @@ get_u32(void)
   return x;
 }
 
+static unsigned long long int
+get_u64(void)
+{
+  uint64_t x;
+  memcpy(&x, &data[pos], 8);
+  pos += 8;
+  return x;
+}
+
 static unsigned int
 get_be32(void)
 {
@@ -120,6 +130,27 @@ match_u32_assert(uint32_t x, const char *where)
 }
 #define match_u32_assert(x) match_u32_assert(x, WHERE)
 
+static bool __attribute__((unused))
+match_u64(uint64_t x)
+{
+  if (get_u64() == x)
+    return true;
+  pos -= 8;
+  return false;
+}
+
+static void __attribute__((unused))
+match_u64_assert(uint64_t x, const char *where)
+{
+  unsigned long long int y = get_u64();
+  if (x != y)
+    {
+      fprintf(stderr, "%s: 0x%x: expected u64:%llu, got u64:%llu\n", where, pos - 8, x, y);
+      exit(1);
+    }
+}
+#define match_u64_assert(x) match_u64_assert(x, WHERE)
+
 static bool __attribute__((unused))
 match_be32(uint32_t x)
 {
@@ -310,18 +341,17 @@ hex_dump(int ofs, int n)
       int c = data[ofs + i];
 #if 1
       if (i && !(i % 16))
-        fprintf(stderr, "-");
+        printf("-");
       else
-        fprintf(stderr, " ");
+        printf(" ");
 #endif
-      fprintf(stderr, "%02x", c);
+      printf("%02x", c);
     }
   for (int i = 0; i < n; i++)
     {
       int c = data[ofs + i];
-      fprintf(stderr, "%c", c >= 32 && c < 127 ? c : '.');
+      printf("%c", c >= 32 && c < 127 ? c : '.');
     }
-  fprintf(stderr, "\n");
 }
 
 static char *
@@ -986,7 +1016,7 @@ dump_fonts(void)
     }
   bool grid = get_byte();
   pos += 3;
-  printf("  <grid show=\"%s\">\n", grid ? "yes" : "no");
+  printf("  <grid show=\"%s\"/>\n", grid ? "yes" : "no");
   printf("</borders>\n");
   assert(pos == x1_end);
 
@@ -1010,7 +1040,11 @@ dump_fonts(void)
       if (!get_byte())
         printf(" footnoteposition=\"subscript\"");
       get_byte();
-      pos += get_be32();
+      int nbytes = get_be32();
+      printf("\n");
+      hex_dump(pos, nbytes);
+      printf("\n");
+      pos += nbytes;
       get_string_be();
       char *look = get_string_be();
       if (look[0])
@@ -1019,8 +1053,16 @@ dump_fonts(void)
     }
   pos = x3_end;
 
+  /* Manual column widths, if present. */
   int count = get_u32();
-  pos += 4 * count;
+  printf("<columnwidths>");
+  for (int i = 0; i < count; i++)
+    {
+      if (i)
+        putchar(' ');
+      printf("%d", get_u32());
+    }
+  printf("</columnwidths>\n");
 
   const char *locale = get_string();
   printf ("<locale>%s</locale>\n", locale);
@@ -1031,19 +1073,7 @@ dump_fonts(void)
   match_byte_assert(0);
   if (!match_byte(0))
     match_byte_assert(1);
-  if (version > 1)
-    {
-      if (!match_byte(0x97) && !match_byte(0x98)
-          && !match_byte(0x99) && !match_byte(0x9a) && !match_byte(0x9b))
-        match_byte_assert(0x9c);
-      match_byte_assert(7);
-      match_byte_assert(0);
-      match_byte_assert(0);
-    }
-  else
-    {
-      printf("%x\n", get_u32());
-    }
+  printf("<epoch>%d</epoch>\n", get_u32());
 
   int decimal = data[pos];
   int grouping = data[pos + 1];
@@ -1078,7 +1108,35 @@ dump_fonts(void)
       int outer_end = get_end();
 
       /* First inner envelope: byte*33 int[n] int*[n]. */
-      pos = get_end();
+      int inner_len = get_u32();
+      int inner_end = pos + inner_len;
+      int array_start = pos + 33;
+      match_byte_assert(0);
+      pos++;                    /* 0, 1, 10 seen. */
+      match_byte_assert(0);
+      pos++;                    /* 0...11 seen. */
+      if (!match_byte(0) && !match_byte(1) && !match_byte(2))
+        match_byte_assert(3);
+      if (!match_byte(0) && !match_byte(2))
+        match_byte_assert(3);
+      if (!match_u64(0))
+        match_u64_assert(UINT64_MAX);
+      match_u32_assert(0);
+      match_u32_assert(0);
+      match_u32_assert(0);
+      match_u32_assert(0);
+      match_byte_assert(0);
+      if (!match_byte(0))
+        match_byte_assert(1);
+      match_byte_assert(1);
+      pos = array_start;
+#if 1
+      printf("widths:");
+      while (pos < inner_end)
+        printf(" %d", get_u32());
+      printf("\n");
+#endif
+      pos = inner_end;;
 
       /* Second inner envelope. */
       assert(get_end() == outer_end);
@@ -1105,12 +1163,7 @@ dump_fonts(void)
       if (!match_byte(0))
         match_byte_assert(1);
 
-      if (!match_byte(0x97) && !match_byte(0x98)
-          && !match_byte(0x99) && !match_byte(0x9a) && !match_byte(0x9b))
-        match_byte_assert(0x9c);
-      match_byte_assert(7);
-      match_byte_assert(0);
-      match_byte_assert(0);
+      printf("<epoch2>%d</epoch2>\n", get_u32());
 
       if (match_byte('.'))
         {
@@ -1124,9 +1177,9 @@ dump_fonts(void)
             match_byte_assert(0);
         }
 
-      pos += 8;
-      match_byte_assert(1);
+      printf ("small: %g\n", get_double());
 
+      match_byte_assert(1);
       if (outer_end - pos > 6)
         {
           /* There might be a pair of strings representing a dataset and
@@ -1146,7 +1199,13 @@ dump_fonts(void)
               printf("<datafile>%s</datafile>\n", get_string());
 
               match_u32_assert(0);
-              get_u32();
+
+              time_t date = get_u32();
+              struct tm tm = *localtime(&date);
+              char s[128];
+              strftime(s, sizeof s, "%a, %d %b %Y %H:%M:%S %z", &tm);
+              printf("<date>%s</date>\n", s);
+
               match_u32_assert(0);
             }
         }
@@ -1159,7 +1218,7 @@ dump_fonts(void)
       else
         match_u32_assert(0);
 
-      match_byte_assert(0x2e);
+      match_byte_assert('.');
       if (!match_byte(0))
         match_byte_assert(1);
 
@@ -1302,19 +1361,20 @@ main(int argc, char *argv[])
       match_byte_assert(0);
 
       /* Offset 27. */
+#if 1
       pos++;
       pos++;
       pos++;
       pos++;
+#else
+      printf("\nheader:");
+      for (int i = 0; i < 4; i++)
+        printf(" %02x", data[pos++]);
+      printf("\n");
+#endif
 
-      /* Offset 31.
-
-         This is the tableId, e.g. -4154297861994971133 would be 0xdca00003.
-         We don't have enough context to validate it. */
-      pos += 4;
-
-      /* Offset 35. */
-      pos += 4;
+      /* Offset 31. */
+      printf("<tableid>%lld</tableid>", get_u64());
 
       dump_title ();
       dump_fonts();