1 files changed, 81 insertions, 0 deletions

diff --git a/src/postprocess/join.c b/src/postprocess/join.c
new file mode 100644
index 0000000..9782ec8
--- /dev/null
+++ b/src/postprocess/join.c
@@ -0,0 +1,81 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include "mystdlib.h"
+
+#define HUGEBLOCK (1024*1024*256)
+#define ZIP_FIELD 3
+#define STREET_FIELD 5
+
+int rt_strcmp( uint8_t *a, uint8_t *b ) {
+  while( ( *a != '\n' ) && ( *b != '\n' ) && ( *a == *b ) ) ++a, ++b;
+  if( *a == *b ) return 0;
+  return -1; 
+}
+
+size_t rt_strcpy( uint8_t *dest, uint8_t *src ) {
+  uint8_t *d = dest;
+  while( *src != '\n' )
+    *dest++ = *src++;
+  *dest++ = '\n';
+  return dest - d;
+} 
+
+size_t rt_strlen( uint8_t *str ) {
+  uint8_t *s = str;
+  while( *str++ != '\n' );
+  return str - s;
+}
+
+int main( int argc, char **argv ) {
+  MAP file = map_file( argv[1], 1 );
+  uint8_t *out, *in;
+  size_t last = 3, off = 0, out_off = 0;
+  int start, end, copy;
+
+  (void)argc;
+
+  out = malloc( HUGEBLOCK );
+
+  if( !file || !out )
+    exit(1);
+
+  in = file->addr;
+  start = 10 * ( in[off] - '0' ) + in[off+1] - '0'; 
+  end = start - 1;
+
+  while( off < file->size ) {
+    int issue = 10 * ( in[off] - '0' ) + in[off+1] - '0'; 
+    off += 3;
+    copy = 1;
+
+//    fprintf( stderr, "issue: %02d start %02d end %02d last %08d off %08d", issue, start, end, last, off );
+    switch ( rt_strcmp( in + last, in + off ) ) {
+    case 1:
+      last = off;
+    case 0:
+    case 2:
+      if (issue == end + 1 ) copy = 0, end++;
+      if (issue == end ) copy = 0;
+      break;
+    default:
+      break;
+    }
+//    fprintf( stderr, " copy: %d\n", copy ); 
+
+    if( copy) {
+      out_off += sprintf( (char*)out + out_off, "%02d%02d\a", start, end );
+      out_off += rt_strcpy( out + out_off, in + last );
+      start = issue; end = issue;
+      last = off;
+    }
+
+    off += rt_strlen( in + off );
+
+    if( out_off + 8192 * 2 > HUGEBLOCK ) {
+      fwrite( out, out_off, 1, stdout );
+      out_off = 0;
+    }
+  }
+
+  return 0;
+}