Resouce name can be terminated with either white-space or colon.

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

diff --git a/index/extract.c b/index/extract.c
index 7af7190..7374470 100644 (file)
--- a/index/extract.c
+++ b/index/extract.c
@@ -1,10 +1,38 @@
 /*
- * Copyright (C) 1994-1996, Index Data I/S 
+ * Copyright (C) 1994-1997, Index Data I/S 
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: extract.c,v $
- * Revision 1.69  1997-04-29 09:26:03  adam
+ * Revision 1.76  1997-10-27 14:33:04  adam
+ * Moved towards generic character mapping depending on "structure"
+ * field in abstract syntax file. Fixed a few memory leaks. Fixed
+ * bug with negative integers when doing searches with relational
+ * operators.
+ *
+ * Revision 1.75  1997/09/17 12:19:12  adam
+ * Zebra version corresponds to YAZ version 1.4.
+ * Changed Zebra server so that it doesn't depend on global common_resource.
+ *
+ * Revision 1.74  1997/09/09 13:38:06  adam
+ * Partial port to WIN95/NT.
+ *
+ * Revision 1.73  1997/09/04 13:57:20  adam
+ * New file extract/retrieve method tellf (added).
+ * Added O_BINARY for open calls.
+ *
+ * Revision 1.72  1997/07/15 16:32:29  adam
+ * Bug fix: Match handler didn't terminate the resulting string!
+ *
+ * Revision 1.71  1997/07/15 16:28:41  adam
+ * Bug fix: storeData didn't work with files with multiple records.
+ * Bug fix: fixed memory management with records; not really well
+ *  thought through.
+ *
+ * Revision 1.70  1997/07/01 13:00:42  adam
+ * Bug fix in routine searchRecordKey: uninitialized variables.
+ *
+ * Revision 1.69  1997/04/29 09:26:03  adam
  * Bug fix: generic recordId handling didn't work for compressed internal
  * keys.
  *
@@ -251,7 +279,11 @@
  */
 #include <stdio.h>
 #include <assert.h>
+#ifdef WINDOWS
+#include <io.h>
+#else
 #include <unistd.h>
+#endif
 #include <fcntl.h>
 
 #include <recctrl.h>
@@ -288,7 +320,7 @@ static void logRecord (int showFlag)
     }
 }
 
-void key_open (int mem)
+void key_open (BFiles bfs, int mem)
 {
     if (!mem)
         mem = atoi(res_get_def (common_resource, "memMax", "4"))*1024*1024;
@@ -301,13 +333,13 @@ void key_open (int mem)
     key_buf_used = 0;
     key_file_no = 0;
 
-    if (!(matchDict = dict_open (GMATCH_DICT, 50, 1)))
+    if (!(matchDict = dict_open (bfs, GMATCH_DICT, 50, 1)))
     {
         logf (LOG_FATAL, "dict_open fail of %s", GMATCH_DICT);
         exit (1);
     }
     assert (!records);
-    records = rec_open (1);
+    records = rec_open (bfs, 1);
 #if 1
     zti = zebTargetInfo_open (records, 1);
 #endif
@@ -481,7 +513,7 @@ void key_flush (void)
     key_buf_used = 0;
 }
 
-int key_close (void)
+int key_close ()
 {
     key_flush ();
     xfree (key_buf);
@@ -499,7 +531,7 @@ static void wordInit (RecWord *p)
 {
     p->attrSet = 1;
     p->attrUse = 1016;
-    p->which = Word_String;
+    p->reg_type = 'w';
 }
 
 struct recKeys {
@@ -563,19 +595,9 @@ static void addRecordKey (const RecWord *p)
         memcpy (dst, &attrUse, sizeof(attrUse));
         dst += sizeof(attrUse);
     }
-    switch (p->which)
-    {
-        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 < IT_MAX_WORD-3; i++)
-        *dst++ = p->u.string[i];
+    *dst++ = p->reg_type;
+    for (i = 0; p->string[i] && i < IT_MAX_WORD-3; i++)
+        *dst++ = p->string[i];
     *dst++ = '\0';
 
     if (!diff)
@@ -656,6 +678,9 @@ static const char **searchRecordKey (struct recKeys *reckeys,
     int off = 0;
     int startSeq = -1;
     int i;
+    short attrUse;
+    char attrSet;
+    int seqno = 0;
 
     for (i = 0; i<32; i++)
         ws[i] = NULL;
@@ -666,9 +691,6 @@ static const char **searchRecordKey (struct recKeys *reckeys,
         const char *src = reckeys->buf + off;
        const char *wstart;
         int lead;
-       short attrUse;
-       char attrSet;
-       int seqno;
     
         lead = *src++;
 
@@ -741,6 +763,12 @@ static off_t file_seek (void *handle, off_t offset)
     return lseek (p->fd, offset, SEEK_SET);
 }
 
+static off_t file_tell (void *handle)
+{
+    struct file_read_info *p = handle;
+    return p->file_offset;
+}
+
 static int file_read (void *handle, char *buf, size_t count)
 {
     struct file_read_info *p = handle;
@@ -915,6 +943,7 @@ static char *fileMatchStr (struct recKeys *reckeys, struct recordGroup *rGroup,
               fname, rGroup->groupName ? rGroup->groupName : "none");
         return NULL;
     }
+    *dst = '\0';
     return dstBuf;
 }
 
@@ -956,6 +985,7 @@ static int recordExtract (SYSNO *sysno, const char *fname,
         extractCtrl.subType = subType;
         extractCtrl.init = wordInit;
         extractCtrl.add = addRecordKey;
+       extractCtrl.dh = rGroup->dh;
 
         reckeys.buf_used = 0;
         reckeys.prevAttrUse = -1;
@@ -966,8 +996,9 @@ static int recordExtract (SYSNO *sysno, const char *fname,
         extractCtrl.offset = recordOffset;
         extractCtrl.readf = file_read;
         extractCtrl.seekf = file_seek;
+        extractCtrl.tellf = file_tell;
         extractCtrl.endf = file_end;
-        extractCtrl.map_chrs_input = map_chrs_input;
+       extractCtrl.zebra_maps = rGroup->zebra_maps;
         extractCtrl.flagShowRecords = rGroup->flagShowRecords;
         if (rGroup->flagShowRecords)
             printf ("File: %s %ld\n", fname, (long) recordOffset);
@@ -1140,16 +1171,18 @@ static int recordExtract (SYSNO *sysno, const char *fname,
     xfree (rec->info[recInfo_storeData]);
     if (rGroup->flagStoreData == 1)
     {
-        rec->size[recInfo_storeData] = fi->file_max;
-        rec->info[recInfo_storeData] = xmalloc (fi->file_max);
+        int size = fi->file_moffset - recordOffset; 
+        if (!size)
+            size = fi->file_max - recordOffset;
+        rec->size[recInfo_storeData] = size;
+        rec->info[recInfo_storeData] = xmalloc (size);
         if (lseek (fi->fd, recordOffset, SEEK_SET) < 0)
         {
             logf (LOG_ERRNO|LOG_FATAL, "seek to %ld in %s", fname,
                   (long) recordOffset);
             exit (1);
         }
-        if (read (fi->fd, rec->info[recInfo_storeData], fi->file_max)
-            < fi->file_max)
+        if (read (fi->fd, rec->info[recInfo_storeData], size) < size)
         {
             logf (LOG_ERRNO|LOG_FATAL, "read %d bytes of %s",
                   fi->file_max, fname);
@@ -1295,7 +1328,7 @@ int fileExtract (SYSNO *sysno, const char *fname,
         fd = -1;
     else
     {
-        if ((fd = open (fname, O_RDONLY)) == -1)
+        if ((fd = open (fname, O_BINARY|O_RDONLY)) == -1)
         {
             logf (LOG_WARN|LOG_ERRNO, "open %s", fname);
             return 0;