Charmap work

[idzebra-moved-to-github.git] / index / extract.c

diff --git a/index/extract.c b/index/extract.c
index 62a7613..c6897dc 100644 (file)
--- a/index/extract.c
+++ b/index/extract.c
@@ -1,10 +1,19 @@
 /*
- * Copyright (C) 1994-1995, Index Data I/S 
+ * Copyright (C) 1994-1996, Index Data I/S 
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: extract.c,v $
- * Revision 1.57  1996-05-13 14:23:04  adam
+ * Revision 1.60  1996-06-04 10:18:12  adam
+ * Search/scan uses character mapping module.
+ *
+ * Revision 1.59  1996/05/14  15:47:07  adam
+ * Cleanup of various buffer size entities.
+ *
+ * Revision 1.58  1996/05/14  06:16:38  adam
+ * Compact use/set bytes used in search service.
+ *
+ * Revision 1.57  1996/05/13 14:23:04  adam
  * Work on compaction of set/use bytes in dictionary.
  *
  * Revision 1.56  1996/05/09  09:54:42  adam
@@ -210,7 +219,6 @@
 #include <assert.h>
 #include <unistd.h>
 #include <fcntl.h>
-#include <ctype.h>
 
 #include <alexutil.h>
 #include <recctrl.h>
@@ -273,7 +281,7 @@ void key_open (int mem)
 struct encode_info {
     int  sysno;
     int  seqno;
-    char buf[512];
+    char buf[768];
 };
 
 void encode_key_init (struct encode_info *i)
@@ -437,32 +445,35 @@ static void addRecordKey (const RecWord *p)
     else
         reckeys.prevAttrUse = attrUse;
 
+    *dst++ = lead;
+
+    if (!(lead & 1))
+    {
+        memcpy (dst, &attrSet, sizeof(attrSet));
+        dst += sizeof(attrSet);
+    }
+    if (!(lead & 2))
+    {
+        memcpy (dst, &attrUse, sizeof(attrUse));
+        dst += sizeof(attrUse);
+    }
     switch (p->which)
     {
-    case Word_String: case Word_Phrase:
-        *dst++ = lead;
-
-        if (!(lead & 1))
-        {
-            memcpy (dst, &attrSet, sizeof(attrSet));
-            dst += sizeof(attrSet);
-        }
-        if (!(lead & 2))
-        {
-            memcpy (dst, &attrUse, sizeof(attrUse));
-            dst += sizeof(attrUse);
-        }
-        for (i = 0; p->u.string[i]; i++)
-            *dst++ = p->u.string[i];
-        *dst++ = '\0';
-
-        memcpy (dst, &p->seqno, sizeof(p->seqno));
-        dst += sizeof(p->seqno);
-
-        break;
-    default:
-        return;
+        case Word_String:
+            *dst++ = 'w';
+            break;
+        case Word_Phrase:
+            *dst++ = 'p';
+            break;
+        case Word_Numeric:
+            *dst++ = 'n';
     }
+    for (i = 0; p->u.string[i]; i++)
+        *dst++ = p->u.string[i];
+    *dst++ = '\0';
+
+    memcpy (dst, &p->seqno, sizeof(p->seqno));
+    dst += sizeof(p->seqno);
     reckeys.buf_used = dst - reckeys.buf;
 }
 
@@ -473,13 +484,11 @@ static void flushRecordKeys (SYSNO sysno, int cmd, struct recKeys *reckeys,
     short attrUse = -1;
     int off = 0;
 
-#if 1
     if (zebTargetInfo_curDatabase (zti, databaseName))
     {
         if (zebTargetInfo_newDatabase (zti, databaseName))
             abort ();
     }
-#endif
     while (off < reckeys->buf_used)
     {
         const char *src = reckeys->buf + off;
@@ -502,18 +511,14 @@ static void flushRecordKeys (SYSNO sysno, int cmd, struct recKeys *reckeys,
             key_flush ();
         ++ptr_i;
         key_buf[ptr_top-ptr_i] = (char*)key_buf + key_buf_used;
-#if 1
+
         lead = zebTargetInfo_lookupSU (zti, attrSet, attrUse);
         if (lead < 0)
             lead = zebTargetInfo_addSU (zti, attrSet, attrUse);
         assert (lead > 0);
         ((char*) key_buf) [key_buf_used++] = lead;
-#else
-        key_buf_used += index_word_prefix ((char*)key_buf + key_buf_used,
-                                           attrSet, attrUse, databaseName);
-#endif
         while (*src)
-            ((char*)key_buf) [key_buf_used++] = index_char_cvt (*src++);
+            ((char*)key_buf) [key_buf_used++] = *src++;
         src++;
         ((char*)key_buf) [key_buf_used++] = '\0';