From 8af10bb39253b97589c5f4b455b708c8fb9e233b Mon Sep 17 00:00:00 2001
From: Ben Pfaff <blp@cs.stanford.edu>
Date: Tue, 5 Jan 2016 22:59:21 -0800
Subject: [PATCH] Figure out a few more details of one of the chunks in the
 fonts.

---
 dump.c | 138 +++++++++++++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 130 insertions(+), 8 deletions(-)

diff --git a/dump.c b/dump.c
index c247c4af66..04f38e0a75 100644
--- a/dump.c
+++ b/dump.c
@@ -218,11 +218,18 @@ get_string(const char *where)
 }
 #define get_string() get_string(WHERE)
 
+static int
+get_end(void)
+{
+  int len = get_u32();
+  return pos + len;
+}
+
 static char *
 dump_counted_string(void)
 {
   char *s = NULL;
-  int inner_end = pos + get_u32();
+  int inner_end = get_end();
   if (pos != inner_end)
     {
       match_u32_assert(0);
@@ -246,7 +253,7 @@ dump_nested_string(void)
 
   match_byte_assert (0);
   match_byte_assert (0);
-  int outer_end = pos + get_u32();
+  int outer_end = get_end();
   s = dump_counted_string();
   match_byte_assert(0x58);
   match_byte_assert(0x58);
@@ -294,8 +301,8 @@ dump_optional_value(FILE *stream)
               return;
             }
 
-          int outer_end = pos + get_u32();
-
+          int outer_end = get_end();
+          
           /* This counted-string appears to be a template string,
              e.g. "Design\: [:^1:]1 Within Subjects Design\: [:^1:]2". */
           dump_counted_string();
@@ -785,6 +792,24 @@ dump_title(void)
     }
 }
 
+static void __attribute__((unused))
+hex_dump(int ofs, int n)
+{
+  for (int i = 0; i < n; i++)
+    {
+      int c = data[ofs + i];
+#if 0
+      if (i && !(i % 16))
+        fprintf(stderr, "-");
+      else
+        fprintf(stderr, " ");
+#endif
+      fprintf(stderr, "%02x ", c);
+      //fprintf(stderr, "%c", c >= 32 && c < 127 ? c : '.');
+    }
+  fprintf(stderr, "\n");
+}
+
 static void
 dump_fonts(void)
 {
@@ -871,8 +896,8 @@ dump_fonts(void)
   int count = get_u32();
   pos += 4 * count;
 
-  const char *encoding = get_string();
-  printf ("<encoding>%s</encoding>\n", encoding);
+  const char *locale = get_string();
+  printf ("<locale>%s</locale>\n", locale);
 
   if (!match_u32(0))
     match_u32_assert(UINT32_MAX);
@@ -916,8 +941,105 @@ dump_fonts(void)
     }
   else
     match_u32_assert(0);
-  int skip = get_u32();
-  pos += skip;
+
+  /* The last chunk is an outer envelope that contains two inner envelopes.
+     The second inner envelope has some interesting data like the encoding and
+     the locale. */
+  if (version == 3)
+    {
+      int outer_end = get_end();
+
+      /* Skip first inner envelope. */
+      int inner_len = get_u32();
+      hex_dump(pos, inner_len);
+      pos += inner_len;
+
+      assert(get_end() == outer_end);
+
+      match_byte_assert(1);
+      match_byte_assert(0);
+      if (!match_byte(3))
+        match_byte_assert(4);
+      match_byte_assert(0);
+      match_byte_assert(0);
+      match_byte_assert(0);
+
+      printf("<command>%s</command>\n", get_string());
+      printf("<subcommand>%s</subcommand>\n", get_string());
+      printf("<language>%s</language>\n", get_string());
+      printf("<charset>%s</charset>\n", get_string());
+      printf("<locale>%s</locale>\n", get_string());
+
+      if (!match_byte(0))
+        match_byte_assert(1);
+      match_byte_assert(0);
+      if (!match_byte(0))
+        match_byte_assert(1);
+      if (!match_byte(0))
+        match_byte_assert(1);
+
+      if (!match_byte(0x97) && !match_byte(0x98) && !match_byte(0x99))
+        match_byte_assert(0x9a);
+      match_byte_assert(7);
+      match_byte_assert(0);
+      match_byte_assert(0);
+
+      if (match_byte('.'))
+        {
+          if (!match_byte(',') && !match_byte('\''))
+            match_byte_assert(' ');
+        }
+      else
+        {
+          match_byte_assert(',');
+          if (!match_byte('.') && !match_byte(' '))
+            match_byte_assert(0);
+        }
+
+      pos += 8;
+      match_byte_assert(1);
+
+      if (outer_end - pos > 6)
+        {
+          uint8_t *endp = memmem(&data[pos], outer_end - pos, "\5\0\0\0\4", 5);
+          assert(endp);
+
+          int end = endp - data;
+          if (pos != end)
+            {
+              printf("<dataset>%s</dataset>\n", get_string());
+              printf("<datafile>%s</datafile>\n", get_string());
+
+              if (pos != end)
+                {
+                  match_u32_assert(0);
+                  get_u32();
+                  match_u32_assert(0);
+                }
+            }
+          //fprintf(stderr, "%d\n", end - pos);
+          //hex_dump(pos, end - pos);
+          assert(pos == end);
+        }
+
+      if (match_u32(5))
+        {
+          for (int i = 0; i < 5; i++)
+            printf("<CC%c>%s</CC%c>\n", 'A' + i, get_string(), 'A' + i);
+        }
+      else
+        match_u32_assert(0);
+
+      match_byte_assert(0x2e);
+      if (!match_byte(0))
+        match_byte_assert(1);
+
+      assert(pos == outer_end);
+
+      pos = outer_end;
+    }
+  else
+    pos = get_end();
 }
 
 int
-- 
2.30.2