Added a new 'cut' directive to charmaps (.chr files) which specifies that

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

diff --git a/index/extract.c b/index/extract.c
index 76c2182..593c9e8 100644 (file)
--- a/index/extract.c
+++ b/index/extract.c
@@ -1,4 +1,4 @@
-/* $Id: extract.c,v 1.152 2004-06-01 12:56:14 adam Exp $
+/* $Id: extract.c,v 1.157 2004-07-28 09:47:41 adam Exp $
    Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004
    Index Data Aps
 
@@ -756,6 +756,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname,
     char ext[128];
     char ext_res[128];
     struct file_read_info *fi;
+    const char *original_record_type = 0;
 
     if (!zh->m_group || !*zh->m_group)
         *gprefix = '\0';
@@ -775,6 +776,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname,
             break;
         }
     /* determine file type - depending on extension */
+    original_record_type = zh->m_record_type;
     if (!zh->m_record_type)
     {
         sprintf (ext_res, "%srecordType.%s", gprefix, ext);
@@ -812,6 +814,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname,
         if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1)
         {
             logf (LOG_WARN|LOG_ERRNO, "open %s", full_rep);
+           zh->m_record_type = original_record_type;
             return 0;
         }
     }
@@ -824,6 +827,7 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname,
     file_read_stop (fi);
     if (fd != -1)
         close (fd);
+    zh->m_record_type = original_record_type;
     return r;
 }
 
@@ -1028,7 +1032,7 @@ int buffer_extract_record (ZebraHandle zh,
                extract_flushSortKeys (zh, *sysno, -1, &zh->reg->sortKeys);
                rec_rm (&rec);
                logRecord(zh);
-               return 0;
+               return -1;
            }
        }
 
@@ -1284,9 +1288,9 @@ void extract_flushRecordKeys (ZebraHandle zh, SYSNO sysno,
         if (zh->reg->key_buf_used + 1024 > 
             (zh->reg->ptr_top -zh->reg->ptr_i)*sizeof(char*))
             extract_flushWriteKeys (zh,0);
-       assert(zh->reg->ptr_i >= 0);
+        assert(zh->reg->ptr_i >= 0);
         ++(zh->reg->ptr_i);
-       assert(zh->reg->ptr_i > 0);
+        assert(zh->reg->ptr_i > 0);
         (zh->reg->key_buf)[zh->reg->ptr_top - zh->reg->ptr_i] =
            (char*)zh->reg->key_buf + zh->reg->key_buf_used;
 #if SU_SCHEME
@@ -1481,6 +1485,10 @@ void extract_add_index_string (RecWord *p, const char *string,
     }
     dst = keys->buf + keys->buf_used;
 
+    /* leader byte is encoded as follows:
+       bit 0 : 1 if attrset is unchanged; 0 if attrset is changed
+       bit 1 : 1 if attruse is unchanged; 0 if attruse is changed
+    */
     attrSet = p->attrSet;
     if (keys->buf_used > 0 && keys->prevAttrSet == attrSet)
         lead |= 1;
@@ -1647,6 +1655,7 @@ static void extract_add_complete_field (RecWord *p)
        while (map && *map && **map == *CHR_SPACE)
        {
            remain = p->length - (b - p->string);
+
            if (remain > 0)
                map = zebra_maps_input(p->zebra_maps, p->reg_type, &b, remain);
            else
@@ -1661,10 +1670,17 @@ static void extract_add_complete_field (RecWord *p)
        {
            const char *cp = *map;
 
-           if (i >= IT_MAX_WORD)
-               break;
-           while (i < IT_MAX_WORD && *cp)
-               buf[i++] = *(cp++);
+           if (**map == *CHR_CUT)
+           {
+               i = 0;
+           }
+           else
+           {
+               if (i >= IT_MAX_WORD)
+                   break;
+               while (i < IT_MAX_WORD && *cp)
+                   buf[i++] = *(cp++);
+           }
            remain = p->length  - (b - p->string);
            if (remain > 0)
                map = zebra_maps_input (p->zebra_maps, p->reg_type, &b,