From 0aa86c5517397a635ca1e63747dd1eefad944960 Mon Sep 17 00:00:00 2001
From: Dirk Engling <erdgeist@erdgeist.org>
Date: Tue, 12 May 2015 03:53:03 +0200
Subject: Split continuation records into files, too.

---
 src/export/split_version_2.c | 98 +++++++++++++++++++++++++++++++++-----------
 1 file changed, 73 insertions(+), 25 deletions(-)

(limited to 'src')

diff --git a/src/export/split_version_2.c b/src/export/split_version_2.c
index 7a6f04e..3ed1bbd 100644
--- a/src/export/split_version_2.c
+++ b/src/export/split_version_2.c
@@ -6,7 +6,56 @@
 
 #include "mystdlib.h"
 
+/*
+
+01_unknown => 01_Flags
+17_unknown => 02_Nachname
+08_unknown => 03_Vorname
+12_unknown => 04_Zusaetze
+  14_unknown => 05_Zusaetze     (sw)
+  15_unknown => 05_Zusaetze     (no)
+13_unknown => 06_Ortszusatz
+09_unknown => 07_Strasse
+11_unknown => 08_Hausnummer
+  15_unknown => 09_Verweise     (sw)
+  14_unknown => 09_Verweise     (no)
+04_unknown => 10_Postleitzahl
+10_unknown => 11_Ort
+  06_unknown => 12_Vorwahl      (sw)
+  07_unknown => 12_Vorwahl      (no)
+  07_unknown => 13_Rufnummer    (sw)
+  06_unknown => 13_Rufnummer    (no)
+
+       no    s     w
+4003   01    01    01
+4007   -     -     5:"0"
+4009   -     *:"0" -
+400B   -     -     6:"0"
+400C   -     !!!   -      => in 1996_Q1: 04_Namenszusatz
+400E   15    07    -
+400F   06    06    14
+4010   -     -     07
+4011   17    -     17
+4012   07    17    06
+4013   14    09    09
+4014   13    11    11
+4015   09    15    04
+4016   11    04    10
+4017   04    10    13
+4018   10    14    -
+401A   -     13    15
+401B   08    08    08
+401C   12    12    12
+
+*/
+
 enum { g_outfiles = 17 };
+static enum { CD_NO = 0, CD_S = 1, CD_W = 2 } g_cd;
+static int g_outfilemap[3][29] = {
+  {-1,-1,-1, 0,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,14, 5,-1,16, 6,13,12, 8,10, 3, 9,-1, 7,11 },
+  {-1,-1,-1, 0,-1,-1,-1,-1,-1,-1,-1,-1,11,-1, 6, 5,-1,-1,16, 8,10,14, 3, 9,13,12, 7,11 },
+  {-1,-1,-1, 0,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,13, 6,16, 5, 8,10, 3, 9,12,-1,14, 7,11 }
+};
 
 static struct {
   int      outfiles[g_outfiles];
@@ -14,7 +63,6 @@ static struct {
   size_t   outfill [g_outfiles];
 } g_state;
 
-static int g_northern_version;;
 static uint8_t cp437_to_iso8859_1_table[] = {
   0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
   0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
@@ -35,22 +83,19 @@ static uint8_t cp437_to_iso8859_1_table[] = {
 };
 
 static void dump_string( uint8_t *start, size_t len, int file ) {
-  uint8_t *dest = g_state.outbuf[file] + g_state.outfill[file];
-  g_state.outfill[file] += len;
-  while( len-- )
-    *(dest++) = cp437_to_iso8859_1_table[ *(start++) ];
+  while( len-- && *start )
+    g_state.outbuf[file][g_state.outfill[file]++] = cp437_to_iso8859_1_table[*start++];
 }
 
 static void dump_field( uint8_t **end, int file ) {
   uint8_t len = (*end)[-1];
-  *end -= len + 1;
-  dump_string( *end, len, file);
+  dump_string( *end -= len + 1, len, file);
 }
 
 static void finish_record( ) {
-  int i;
-  for( i=0; i<g_outfiles; ++i )
-    *( g_state.outbuf[i] + g_state.outfill[i]++ ) = '\n';
+  int file;
+  for( file=0; file<g_outfiles; ++file )
+    g_state.outbuf[file][g_state.outfill[file]++] = '\n';
 }
 
 static uint16_t load_word( uint8_t *table_start, uint16_t base, int offset ) {
@@ -59,11 +104,10 @@ static uint16_t load_word( uint8_t *table_start, uint16_t base, int offset ) {
   /* If a word(subflag, t_off) would span over the end of a 0x4000
      byte page, it is placed on the next page. All other words are
      shifted backwards, as well. */
-  table_start += 4;
   if( base + 4 + 4 * dword_offset > 0x1ffc )
     table_start += 3 & ( 4 - ( base & 3 ) );
 
-  return ((uint16_t*)table_start)[offset];
+  return ((uint16_t*)table_start)[offset + 2];
 }
 
 static void dump_primary( uint8_t *end, uint32_t flags ) {
@@ -74,7 +118,7 @@ static void dump_primary( uint8_t *end, uint32_t flags ) {
   dump_string( end -= 5, 5, 1 );
 
   /* Dump first 5 chars of prefix, they're always here in s and w */
-  if( !g_northern_version )
+  if( g_cd != CD_NO )
     dump_string( end -= 5, 5, 2 );
 
   /* There is another version of the zip code present, if this bit is set */
@@ -83,9 +127,8 @@ static void dump_primary( uint8_t *end, uint32_t flags ) {
   /* There is an unclear X present, if this bit is set */
   if( flags & 0x0040 ) dump_string( end -= 1, 1, 4 );
 
-  /* There is another version of the prefix present, if this bit is set */
   if( flags & 0x0020 ) {
-    if( !g_northern_version )
+    if( g_cd != CD_NO )
       dump_string( end -= 5, 5, 5 );
     else
       dump_field( &end, 5 );
@@ -104,7 +147,7 @@ static void dump_primary( uint8_t *end, uint32_t flags ) {
   dump_field( &end, 15 );
 
   /* Dump first 5 chars of prefix, they're always here in no */
-  if( g_northern_version )
+  if( g_cd == CD_NO )
     dump_field( &end, 2 );
   dump_field( &end, 16 );
 }
@@ -117,23 +160,22 @@ static void act_on_record( uint8_t * end, uint16_t base ) {
   uint32_t flags = 0;
   int i;
 
-  for (i=0; i<flagbytes; ++i) flags = (flags<<8) | *--end;
+  while(flagbytes--) flags = (flags<<8) | *--end;
 
 //  printf( "-------- %03d: %06X\n", num_dwords, flags );
   for (i = 0; i < num_dwords; i++ ) {
-    uint16_t subflag = load_word( e, base, 2*i );
-    uint16_t t_off   = load_word( e, base, 2*i+1 );
+    uint16_t subflag   = load_word( e, base, 2*i );
+    uint16_t t_off     = load_word( e, base, 2*i+1 );
     uint8_t *rec_start = end - t_off;
 
     if (i == 0)
       dump_primary( rec_start, flags );
     else {
       uint8_t * rec_end = end - ( ( i+1 == num_dwords ) ? 0 : load_word( e, base, 2*i+3 ) );
-      printf( "(%04X): ", subflag );
-//      dump_string( rec_start, rec_end, file );
-      while ( rec_start < rec_end )
-        putchar( cp437_to_iso8859_1_table[ *(rec_start++) ] );
-      putchar(10);
+      int file = g_outfilemap[g_cd][subflag&0x1F];
+      if( subflag == 0x4003 ) finish_record();
+      if( file >= 0 )
+        dump_string( rec_start, rec_end - rec_start, g_outfilemap[g_cd][subflag&0x1F] );
     }
   }
   finish_record();
@@ -165,7 +207,13 @@ int main( int args, char **argv ) {
   if( !data || !index )
     exit( 1 );
 
-  g_northern_version = data->addr[0x21e] != 5;
+  switch( data->addr[0x228] ) {
+    case 0x07: g_cd = CD_NO; break;
+    case 0x0d: g_cd = CD_S;  break;
+    case 0x0e: g_cd = CD_W;  break;
+    default: fprintf( stderr, "Unknown CD version\n" );
+      exit(1);
+  }
 
   /* Each entry in outfile[0] is flag + \n, i.e. 3 bytes
      We want to flush at ever percent progress */
-- 
cgit v1.2.3