X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=rset%2Frsbool.c;h=b14cc37083e382f3f14c0971e55d977020857235;hb=892728fe00debf2b9176e0bc7517a73113f6660b;hp=bdc39e6b2f20e66f8112a67707d1b8f545a47cb4;hpb=ef696645cc3b7e0f4027008d1dc589c0f0f90c1f;p=idzebra-moved-to-github.git

diff --git a/rset/rsbool.c b/rset/rsbool.c
index bdc39e6..b14cc37 100644
--- a/rset/rsbool.c
+++ b/rset/rsbool.c
@@ -1,68 +1,24 @@
-/*
- * Copyright (C) 1994-1999, Index Data
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: rsbool.c,v $
- * Revision 1.16  1999-05-26 07:49:14  adam
- * C++ compilation.
- *
- * Revision 1.15  1999/02/02 14:51:32  adam
- * Updated WIN32 code specific sections. Changed header.
- *
- * Revision 1.14  1998/03/05 08:36:27  adam
- * New result set model.
- *
- * Revision 1.13  1997/12/18 10:54:24  adam
- * New method result set method rs_hits that returns the number of
- * hits in result-set (if known). The ranked result set returns real
- * number of hits but only when not combined with other operands.
- *
- * Revision 1.12  1997/10/31 12:37:01  adam
- * Code calls xfree() instead of free().
- *
- * Revision 1.11  1997/09/09 13:38:15  adam
- * Partial port to WIN95/NT.
- *
- * Revision 1.10  1996/10/29 13:55:20  adam
- * Include of zebrautl.h instead of alexutil.h.
- *
- * Revision 1.9  1995/12/11 09:15:22  adam
- * New set types: sand/sor/snot - ranked versions of and/or/not in
- * ranked/semi-ranked result sets.
- * Note: the snot not finished yet.
- * New rset member: flag.
- * Bug fix: r_delete in rsrel.c did free bad memory block.
- *
- * Revision 1.8  1995/10/12  12:41:55  adam
- * Private info (buf) moved from struct rset_control to struct rset.
- * Bug fixes in relevance.
- *
- * Revision 1.7  1995/10/10  14:00:03  adam
- * Function rset_open changed its wflag parameter to general flags.
- *
- * Revision 1.6  1995/10/06  14:38:05  adam
- * New result set method: r_score.
- * Local no (sysno) and score is transferred to retrieveCtrl.
- *
- * Revision 1.5  1995/09/08  14:52:41  adam
- * Work on relevance feedback.
- *
- * Revision 1.4  1995/09/08  08:54:04  adam
- * More efficient and operation.
- *
- * Revision 1.3  1995/09/07  13:58:43  adam
- * New parameter: result-set file descriptor (RSFD) to support multiple
- * positions within the same result-set.
- * Boolean operators: and, or, not implemented.
- *
- * Revision 1.2  1995/09/06  16:11:55  adam
- * More work on boolean sets.
- *
- * Revision 1.1  1995/09/06  13:27:15  adam
- * New set type: bool. Not finished yet.
- *
- */
+/* $Id: rsbool.c,v 1.27 2004-06-01 15:22:58 adam Exp $
+   Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004
+   Index Data Aps
+
+This file is part of the Zebra server.
+
+Zebra is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Zebra; see the file LICENSE.zebra.  If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
+*/
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -72,11 +28,21 @@
 #include <rsbool.h>
 #include <zebrautl.h>
 
+/* for key_logdump. Debugging only */
+#include <../index/index.h> 
+
+#ifndef RSET_DEBUG
+#define RSET_DEBUG 0
+#endif
+
 static void *r_create(RSET ct, const struct rset_control *sel, void *parms);
 static RSFD r_open (RSET ct, int flag);
 static void r_close (RSFD rfd);
 static void r_delete (RSET ct);
 static void r_rewind (RSFD rfd);
+static int r_forward(RSET ct, RSFD rfd, void *buf, int *term_index,
+                     int (*cmpfunc)(const void *p1, const void *p2),
+                     const void *untilbuf);
 static int r_count (RSET ct);
 static int r_read_and (RSFD rfd, void *buf, int *term_index);
 static int r_read_or (RSFD rfd, void *buf, int *term_index);
@@ -91,6 +57,7 @@ static const struct rset_control control_and =
     r_close,
     r_delete,
     r_rewind,
+    r_forward, /* rset_default_forward, */
     r_count,
     r_read_and,
     r_write,
@@ -104,6 +71,11 @@ static const struct rset_control control_or =
     r_close,
     r_delete,
     r_rewind,
+#if 1
+    r_forward, 
+#else
+    rset_default_forward,
+#endif
     r_count,
     r_read_or,
     r_write,
@@ -117,6 +89,7 @@ static const struct rset_control control_not =
     r_close,
     r_delete,
     r_rewind,
+    r_forward, 
     r_count,
     r_read_not,
     r_write,
@@ -145,6 +118,7 @@ struct rset_bool_rfd {
     int term_index_r;
     void *buf_l;
     void *buf_r;
+    int tail;
     struct rset_bool_rfd *next;
     struct rset_bool_info *info;
 };    
@@ -188,6 +162,7 @@ static RSFD r_open (RSET ct, int flag)
 	return NULL;
     }
     rfd = (struct rset_bool_rfd *) xmalloc (sizeof(*rfd));
+    logf(LOG_DEBUG,"rsbool (%s) open [%p]", ct->control->desc, rfd);
     rfd->next = info->rfd_list;
     info->rfd_list = rfd;
     rfd->info = info;
@@ -200,6 +175,7 @@ static RSFD r_open (RSET ct, int flag)
 			     &rfd->term_index_l);
     rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r,
 			     &rfd->term_index_r);
+    rfd->tail = 0;
     return rfd;
 }
 
@@ -246,55 +222,254 @@ static void r_rewind (RSFD rfd)
     p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, &p->term_index_r);
 }
 
+static int r_forward (RSET ct, RSFD rfd, void *buf, int *term_index,
+                     int (*cmpfunc)(const void *p1, const void *p2),
+                     const void *untilbuf)
+{
+    struct rset_bool_info *info = ((struct rset_bool_rfd*)rfd)->info;
+    struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd;
+    int cmp=0;
+    int rc;
+
+#if RSET_DEBUG
+    logf (LOG_DEBUG, "rsbool_forward (L) [%p] '%s' (ct=%p rfd=%p m=%d,%d)",
+                      rfd, ct->control->desc, ct, rfd, p->more_l, p->more_r);
+#endif
+    if ( p->more_l && ((cmpfunc)(untilbuf,p->buf_l)==2) )
+        p->more_l = rset_forward(info->rset_l, p->rfd_l, p->buf_l,
+                        &p->term_index_l, info->cmp, untilbuf);
+#if RSET_DEBUG
+    logf (LOG_DEBUG, "rsbool_forward (R) [%p] '%s' (ct=%p rfd=%p m=%d,%d)",
+                      rfd, ct->control->desc, ct, rfd, p->more_l, p->more_r);
+#endif
+    if ( p->more_r && ((cmpfunc)(untilbuf,p->buf_r)==2))
+        p->more_r = rset_forward(info->rset_r, p->rfd_r, p->buf_r,
+                        &p->term_index_r, info->cmp, untilbuf);
+#if RSET_DEBUG
+    logf (LOG_DEBUG, "rsbool_forward [%p] calling read, m=%d,%d t=%d", 
+                       rfd, p->more_l, p->more_r, p->tail);
+#endif
+    
+    p->tail=0; 
+    rc = rset_read(ct,rfd,buf,term_index); 
+#if RSET_DEBUG
+    logf (LOG_DEBUG, "rsbool_forward returning [%p] %d m=%d,%d", 
+                       rfd, rc, p->more_l, p->more_r);
+#endif
+    return rc;
+
+    if (p->more_l && p->more_r)
+           cmp = (*info->cmp)(p->buf_l, p->buf_r);
+       else if (p->more_l)
+           cmp = -2;
+       else
+           cmp = 2;
+    if ( (cmp<0) && (p->more_l) )
+    {
+        memcpy (buf, p->buf_l, info->key_size);
+	    *term_index = p->term_index_l;
+#if RSET_DEBUG
+        logf (LOG_DEBUG, "rsbool_forward returning L (cmp=%d)",cmp);
+#endif
+        return 1;
+    } else if ( (cmp>0) && (p->more_r) )
+    {
+        memcpy (buf, p->buf_r, info->key_size);
+	    *term_index = p->term_index_r + info->term_index_s;
+#if RSET_DEBUG
+        logf (LOG_DEBUG, "rsbool_forward returning R (cmp=%d)",cmp);
+#endif
+        return 1;
+    }
+    /* return ( p->more_l || p->more_r); */
+    return 0;
+}
+
 static int r_count (RSET ct)
 {
     return 0;
 }
 
+
+/*
+    1,1         1,3
+    1,9         2,1
+    1,11        3,1
+    2,9
+
+  1,1     1,1
+  1,3     1,3
+          1,9
+          1,11
+  2,1     2,1
+          2,9
+          3,1
+*/
+
 static int r_read_and (RSFD rfd, void *buf, int *term_index)
 {
     struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd;
     struct rset_bool_info *info = p->info;
 
-    while (p->more_l && p->more_r)
+    while (p->more_l || p->more_r)
     {
         int cmp;
 
-        cmp = (*info->cmp)(p->buf_l, p->buf_r);
+        if (p->more_l && p->more_r)
+            cmp = (*info->cmp)(p->buf_l, p->buf_r);
+        else if (p->more_l)
+            cmp = -2;
+        else
+            cmp = 2;
+#if RSET_DEBUG
+        logf (LOG_DEBUG, "r_read_and [%p] looping: m=%d/%d c=%d t=%d",
+                        rfd, p->more_l, p->more_r, cmp, p->tail);
+        key_logdump(LOG_DEBUG,p->buf_l);
+        key_logdump(LOG_DEBUG,p->buf_r);
+#endif
         if (!cmp)
         {
             memcpy (buf, p->buf_l, info->key_size);
-	    *term_index = p->term_index_l;
+	        *term_index = p->term_index_l;
             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
-            p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
-				   &p->term_index_r);
-            return 1;
+            p->tail = 1;
         }
         else if (cmp == 1)
         {
             memcpy (buf, p->buf_r, info->key_size);
-
-	    *term_index = p->term_index_r + info->term_index_s;
+	        *term_index = p->term_index_r + info->term_index_s;
             p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
 				   &p->term_index_r);
+            p->tail = 1;
+#if RSET_DEBUG
+            logf (LOG_DEBUG, "r_read_and [%p] returning R m=%d/%d c=%d",
+                    rfd, p->more_l, p->more_r, cmp);
+            key_logdump(LOG_DEBUG,buf);
+#endif
             return 1;
         }
         else if (cmp == -1)
         {
             memcpy (buf, p->buf_l, info->key_size);
-	    *term_index = p->term_index_l;
+	        *term_index = p->term_index_l;
             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
+            p->tail = 1;
+#if RSET_DEBUG
+            logf (LOG_DEBUG, "r_read_and [%p] returning L m=%d/%d c=%d",
+                    rfd, p->more_l, p->more_r, cmp);
+            key_logdump(LOG_DEBUG,buf);
+#endif
             return 1;
         }
-        else if (cmp > 1)
+        else if (cmp > 1)  /* cmp == 2 */
+        {
+#define OLDCODE 0
+#if OLDCODE
+            memcpy (buf, p->buf_r, info->key_size);
+            *term_index = p->term_index_r + info->term_index_s;
+            
             p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
-				   &p->term_index_r);
-        else
-            p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
+                                   &p->term_index_r);
+            if (p->tail)
+            {
+                if (!p->more_r || (*info->cmp)(p->buf_r, buf) > 1)
+                    p->tail = 0;
+#if RSET_DEBUG
+                logf (LOG_DEBUG, "r_read_and returning C m=%d/%d c=%d",
+                        p->more_l, p->more_r, cmp);
+                key_logdump(LOG_DEBUG,buf);
+#endif
+                return 1;
+            }
+#else
+            
+            if (p->tail)
+            {
+                memcpy (buf, p->buf_r, info->key_size);
+                *term_index = p->term_index_r + info->term_index_s;
+                p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
+                                   &p->term_index_r);
+                if (!p->more_r || (*info->cmp)(p->buf_r, buf) > 1)
+                    p->tail = 0;
+#if RSET_DEBUG
+                logf (LOG_DEBUG, "r_read_and [%p] returning R tail m=%d/%d c=%d",
+                        rfd, p->more_l, p->more_r, cmp);
+                key_logdump(LOG_DEBUG,buf);
+#endif
+                return 1;
+            } else
+            {
+#if RSET_DEBUG
+                logf (LOG_DEBUG, "r_read_and [%p] about to forward R m=%d/%d c=%d",
+                        rfd, p->more_l, p->more_r, cmp);
+#endif
+                if (p->more_r && p->more_l)
+                    p->more_r = rset_forward( 
+                                    info->rset_r, p->rfd_r, 
+                                    p->buf_r, &p->term_index_r, 
+                                    (info->cmp), p->buf_l);
+                else 
+                    return 0; /* no point in reading further */
+            }
+#endif
+        }
+        else  /* cmp == -2 */
+        {
+#if OLDCODE
+             memcpy (buf, p->buf_l, info->key_size);
+             *term_index = p->term_index_l;
+             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
+                                    &p->term_index_l);
+             if (p->tail)
+             {
+                 if (!p->more_l || (*info->cmp)(p->buf_l, buf) > 1)
+                     p->tail = 0;
+#if RSET_DEBUG
+                 logf (LOG_DEBUG, "r_read_and [%p] returning R tail m=%d/%d c=%d",
+                        rfd, p->more_l, p->more_r, cmp);
+                 key_logdump(LOG_DEBUG,buf);
+#endif
+                 return 1;
+             }
+#else
+            if (p->tail)
+            {
+                memcpy (buf, p->buf_l, info->key_size);
+	            *term_index = p->term_index_l;
+                p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
+                if (!p->more_l || (*info->cmp)(p->buf_l, buf) > 1)
+                    p->tail = 0;
+#if RSET_DEBUG
+                logf (LOG_DEBUG, "r_read_and [%p] returning L tail m=%d/%d c=%d",
+                        rfd, p->more_l, p->more_r, cmp);
+                key_logdump(LOG_DEBUG,buf);
+#endif
+                return 1;
+            }
+            else
+            {
+#if RSET_DEBUG
+                logf (LOG_DEBUG, "r_read_and [%p] about to forward L m=%d/%d c=%d",
+                        rfd, p->more_l, p->more_r, cmp);
+#endif
+                if (p->more_r && p->more_l)
+                    p->more_l = rset_forward( 
+                    /* p->more_l = rset_default_forward( */
+                                    info->rset_l, p->rfd_l, 
+                                    p->buf_l, &p->term_index_l, 
+                                    (info->cmp), p->buf_r);
+                else 
+                    return 0; /* no point in reading further */
+            }
+#endif
+        }
     }
+#if RSET_DEBUG
+    logf (LOG_DEBUG, "r_read_and [%p] reached its end",rfd);
+#endif
     return 0;
 }
 
@@ -316,27 +491,42 @@ static int r_read_or (RSFD rfd, void *buf, int *term_index)
         if (!cmp)
         {
             memcpy (buf, p->buf_l, info->key_size);
-	    *term_index = p->term_index_l;
+	        *term_index = p->term_index_l;
             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
             p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
 				   &p->term_index_r);
+#if RSET_DEBUG
+            logf (LOG_DEBUG, "r_read_or returning A m=%d/%d c=%d",
+                    p->more_l, p->more_r, cmp);
+            key_logdump(LOG_DEBUG,buf);
+#endif
             return 1;
         }
         else if (cmp > 0)
         {
             memcpy (buf, p->buf_r, info->key_size);
-	    *term_index = p->term_index_r + info->term_index_s;
+	        *term_index = p->term_index_r + info->term_index_s;
             p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
 				   &p->term_index_r);
+#if RSET_DEBUG
+            logf (LOG_DEBUG, "r_read_or returning B m=%d/%d c=%d",
+                    p->more_l, p->more_r, cmp);
+            key_logdump(LOG_DEBUG,buf);
+#endif
             return 1;
         }
         else
         {
             memcpy (buf, p->buf_l, info->key_size);
-	    *term_index = p->term_index_l;
+	        *term_index = p->term_index_l;
             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
+#if RSET_DEBUG
+            logf (LOG_DEBUG, "r_read_or returning C m=%d/%d c=%d",
+                    p->more_l, p->more_r, cmp);
+            key_logdump(LOG_DEBUG,buf);
+#endif
             return 1;
         }
     }
@@ -361,19 +551,26 @@ static int r_read_not (RSFD rfd, void *buf, int *term_index)
         if (cmp < -1)
         {
             memcpy (buf, p->buf_l, info->key_size);
-	    *term_index = p->term_index_l;
+	        *term_index = p->term_index_l;
             p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				   &p->term_index_l);
             return 1;
         }
         else if (cmp > 1)
+#if 0
             p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
 				   &p->term_index_r);
+#else
+            p->more_r = rset_forward( 
+                            info->rset_r, p->rfd_r, 
+                            p->buf_r, &p->term_index_r, 
+                            (info->cmp), p->buf_l);
+#endif
         else
         {
             memcpy (buf, p->buf_l, info->key_size);
             do
-            {
+            { 
                 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
 				       &p->term_index_l);
                 if (!p->more_l)