X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fkinput.c;h=c5605914fc382be0a5333aa21f69494b96587210;hb=10e178572346e8c5c3caaa43b803dd10c005cb5f;hp=eb7b251320ec7c39b998a25097211c498a37148d;hpb=2954746ef5c897ec5e651cc2c39eda878f1d5da5;p=idzebra-moved-to-github.git

diff --git a/index/kinput.c b/index/kinput.c
index eb7b251..c560591 100644
--- a/index/kinput.c
+++ b/index/kinput.c
@@ -1,10 +1,42 @@
 /*
- * Copyright (C) 1994-1996, Index Data I/S 
+ * Copyright (C) 1994-1999, Index Data
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: kinput.c,v $
- * Revision 1.23  1997-09-04 13:57:39  adam
+ * Revision 1.33  1999-05-15 14:36:38  adam
+ * Updated dictionary. Implemented "compression" of dictionary.
+ *
+ * Revision 1.32  1999/05/12 13:08:06  adam
+ * First version of ISAMS.
+ *
+ * Revision 1.31  1999/02/02 14:50:56  adam
+ * Updated WIN32 code specific sections. Changed header.
+ *
+ * Revision 1.30  1998/10/28 10:53:57  adam
+ * Added type cast to prevent warning.
+ *
+ * Revision 1.29  1998/06/11 15:41:39  adam
+ * Minor changes.
+ *
+ * Revision 1.28  1998/03/05 08:45:12  adam
+ * New result set model and modular ranking system. Moved towards
+ * descent server API. System information stored as "SGML" records.
+ *
+ * Revision 1.27  1998/02/17 10:32:52  adam
+ * Fixed bug: binary files weren't opened with flag b on NT.
+ *
+ * Revision 1.26  1998/01/29 13:39:13  adam
+ * Compress ISAM is default.
+ *
+ * Revision 1.25  1997/09/17 12:19:14  adam
+ * Zebra version corresponds to YAZ version 1.4.
+ * Changed Zebra server so that it doesn't depend on global common_resource.
+ *
+ * Revision 1.24  1997/09/09 13:38:07  adam
+ * Partial port to WIN95/NT.
+ *
+ * Revision 1.23  1997/09/04 13:57:39  adam
  * Added O_BINARY for open calls.
  *
  * Revision 1.22  1997/02/12 20:39:45  adam
@@ -88,7 +120,11 @@
  */
 
 #include <fcntl.h>
+#ifdef WIN32
+#include <io.h>
+#else
 #include <unistd.h>
+#endif
 #include <stdlib.h>
 #include <string.h>
 #include <stdio.h>
@@ -133,7 +169,7 @@ void getFnameTmp (char *fname, int no)
 
 void key_file_chunk_read (struct key_file *f)
 {
-    int nr = 0, r, fd;
+    int nr = 0, r = 0, fd;
     char fname[1024];
     getFnameTmp (fname, f->no);
     fd = open (fname, O_BINARY|O_RDONLY);
@@ -282,6 +318,7 @@ struct heap_info {
     Dict dict;
     ISAM isam;
     ISAMC isamc;
+    ISAMS isams;
 };
 
 struct heap_info *key_heap_init (int nkeys,
@@ -434,7 +471,7 @@ int heap_inpc (struct heap_info *hi)
         char *dict_info;
 
         strcpy (this_name, hci.cur_name);
-        logf (LOG_DEBUG, "inserting %s", 1+hci.cur_name);
+	assert (hci.cur_name[1]);
         no_diffs++;
         if ((dict_info = dict_lookup (hi->dict, hci.cur_name)))
         {
@@ -465,6 +502,42 @@ int heap_inpc (struct heap_info *hi)
     return 0;
 } 
 
+int heap_inps (struct heap_info *hi)
+{
+    struct heap_cread_info hci;
+    ISAMS_I isams_i = xmalloc (sizeof(*isams_i));
+
+    hci.key = xmalloc (KEY_SIZE);
+    hci.mode = 1;
+    hci.hi = hi;
+    hci.more = heap_read_one (hi, hci.cur_name, hci.key);
+
+    isams_i->clientData = &hci;
+    isams_i->read_item = heap_cread_item;
+
+    while (hci.more)
+    {
+        char this_name[INP_NAME_MAX];
+        ISAMS_P isams_p;
+        char *dict_info;
+
+        strcpy (this_name, hci.cur_name);
+	assert (hci.cur_name[1]);
+        no_diffs++;
+        if (!(dict_info = dict_lookup (hi->dict, hci.cur_name)))
+        {
+            isams_p = isams_merge (hi->isams, isams_i);
+            no_insertions++;
+            dict_insert (hi->dict, this_name, sizeof(ISAMS_P), &isams_p);
+        }
+	else
+	    abort();
+    }
+    xfree (isams_i);
+    return 0;
+} 
+
+
 int heap_inp (struct heap_info *hi)
 {
     char *info;
@@ -491,7 +564,7 @@ int heap_inp (struct heap_info *hi)
                 break;
             memcpy (key_buf + key_buf_ptr, next_key, KEY_SIZE);
             key_buf_ptr += KEY_SIZE;
-            if (key_buf_ptr+KEY_SIZE >= key_buf_size)
+            if (key_buf_ptr+(int) KEY_SIZE >= key_buf_size)
             {
                 char *new_key_buf;
                 new_key_buf = xmalloc (key_buf_size + INP_BUF_ADD);
@@ -503,11 +576,10 @@ int heap_inp (struct heap_info *hi)
         }
         no_diffs++;
         nmemb = key_buf_ptr / KEY_SIZE;
-        assert (nmemb*KEY_SIZE == key_buf_ptr);
+        assert (nmemb * (int) KEY_SIZE == key_buf_ptr);
         if ((info = dict_lookup (hi->dict, cur_name)))
         {
             ISAM_P isam_p, isam_p2;
-            logf (LOG_DEBUG, "updating %s", 1+cur_name);
             memcpy (&isam_p, info+1, sizeof(ISAM_P));
             isam_p2 = is_merge (hi->isam, isam_p, nmemb, key_buf);
             if (!isam_p2)
@@ -526,7 +598,6 @@ int heap_inp (struct heap_info *hi)
         else
         {
             ISAM_P isam_p;
-            logf (LOG_DEBUG, "inserting %s", 1+cur_name);
             no_insertions++;
             isam_p = is_merge (hi->isam, 0, nmemb, key_buf);
             dict_insert (hi->dict, cur_name, sizeof(ISAM_P), &isam_p);
@@ -556,8 +627,8 @@ void progressFunc (struct key_file *keyp, void *info)
     if (now >= p->lastTime+10)
     {
         p->lastTime = now;
-        remaining = (now - p->startTime)*
-            ((double) p->totalBytes/p->totalOffset - 1.0);
+        remaining = (time_t) ((now - p->startTime)*
+            ((double) p->totalBytes/p->totalOffset - 1.0));
         if (remaining <= 130)
             logf (LOG_LOG, "Merge %2.1f%% completed; %ld seconds remaining",
                  (100.0*p->totalOffset) / p->totalBytes, (long) remaining);
@@ -568,12 +639,17 @@ void progressFunc (struct key_file *keyp, void *info)
     p->totalOffset += keyp->buf_size;
 }
 
-void key_input (int nkeys, int cache)
+#ifndef R_OK
+#define R_OK 4
+#endif
+
+void key_input (BFiles bfs, int nkeys, int cache)
                 
 {
     Dict dict;
     ISAM isam = NULL;
     ISAMC isamc = NULL;
+    ISAMS isams = NULL;
     struct key_file **kf;
     char rbuf[1024];
     int i, r;
@@ -594,30 +670,42 @@ void key_input (int nkeys, int cache)
         if (!nkeys)
             return ;
     }
-    dict = dict_open (FNAME_DICT, cache, 1);
+    dict = dict_open (bfs, FNAME_DICT, cache, 1, 0);
     if (!dict)
     {
         logf (LOG_FATAL, "dict_open fail");
         exit (1);
     }
-    if (res_get_match (common_resource, "isam", "c", NULL))
+    if (res_get_match (common_resource, "isam", "s", NULL))
     {
-        isamc = isc_open (FNAME_ISAMC, 1, key_isamc_m ());
-        if (!isamc)
+        isams = isams_open (bfs, FNAME_ISAMS, 1,
+			    key_isams_m (common_resource));
+        if (!isams)
         {
-            logf (LOG_FATAL, "isc_open fail");
+            logf (LOG_FATAL, "isams_open fail");
             exit (1);
         }
     }
-    else
+    else if (res_get_match (common_resource, "isam", "i", NULL))
     {
-        isam = is_open (FNAME_ISAM, key_compare, 1, sizeof(struct it_key));
+        isam = is_open (bfs, FNAME_ISAM, key_compare, 1,
+			sizeof(struct it_key), common_resource);
         if (!isam)
         {
             logf (LOG_FATAL, "is_open fail");
             exit (1);
         }
     }
+    else
+    {
+        isamc = isc_open (bfs, FNAME_ISAMC, 1,
+			  key_isamc_m (common_resource));
+        if (!isamc)
+        {
+            logf (LOG_FATAL, "isc_open fail");
+            exit (1);
+        }
+    }
     kf = xmalloc ((1+nkeys) * sizeof(*kf));
     progressInfo.totalBytes = 0;
     progressInfo.totalOffset = 0;
@@ -635,19 +723,24 @@ void key_input (int nkeys, int cache)
     hi->dict = dict;
     hi->isam = isam;
     hi->isamc = isamc;
+    hi->isams = isams;
 
     for (i = 1; i<=nkeys; i++)
         if ((r = key_file_read (kf[i], rbuf)))
             key_heap_insert (hi, rbuf, r, kf[i]);
     if (isamc)
         heap_inpc (hi);
-    else
-        heap_inp (hi);
+    else if (isams)
+	heap_inps (hi);
+    else if (isam)
+	heap_inp (hi);
     dict_close (dict);
     if (isam)
         is_close (isam);
     if (isamc)
         isc_close (isamc);
+    if (isams)
+	isams_close (isams);
    
     for (i = 1; i<=nkeys; i++)
     {