merge.c

harvest是一个下载html网页得机器人

/* $Id: merge.c,v 1.23 2003/06/23 15:36:11 adam Exp $
   Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
   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 <stdlib.h>
#include <assert.h>
#include <string.h>
#include <stdio.h>
#include <yaz/log.h>
#include "isamc-p.h"

struct isc_merge_block {
    int offset;       /* offset in r_buf */
    int block;        /* block number of file (0 if none) */
    int dirty;        /* block is different from that on file */
};

#if 0
static void opt_blocks (ISAMC is, struct isc_merge_block *mb, int ptr,
			int last)
{
    int i, no_dirty = 0;
    for (i = 0; i<ptr; i++)
	if (mb[i].dirty)
	    no_dirty++;
    if (no_dirty*4 < ptr*3)
	return;
    /* bubble-sort it */
    for (i = 0; i<ptr; i++)
    {
	int tmp, j, j_min = -1;
	for (j = i; j<ptr; j++)
	{
	    if (j_min < 0 || mb[j_min].block > mb[j].block)
		j_min = j;
	}
	assert (j_min >= 0);
	tmp = mb[j_min].block;
	mb[j_min].block = mb[i].block;
	mb[i].block = tmp;
	mb[i].dirty = 1;
    }
    if (!last)
	mb[i].dirty = 1;
}
#endif

static void flush_blocks (ISAMC is, struct isc_merge_block *mb, int ptr,
                          char *r_buf, int *firstpos, int cat, int last,
                          int *numkeys)
{
    int i;

    for (i = 0; i<ptr; i++)
    {
        /* consider this block number */
        if (!mb[i].block) 
        {
            mb[i].block = isc_alloc_block (is, cat);
            mb[i].dirty = 1;
        }

        /* consider next block pointer */
        if (last && i == ptr-1)
            mb[i+1].block = 0;
        else if (!mb[i+1].block)       
        {
            mb[i+1].block = isc_alloc_block (is, cat);
            mb[i+1].dirty = 1;
            mb[i].dirty = 1;
        }
    }

    for (i = 0; i<ptr; i++)
    {
        char *src;
        ISAMC_BLOCK_SIZE ssize = mb[i+1].offset - mb[i].offset;

        assert (ssize);

        /* skip rest if not dirty */
        if (!mb[i].dirty)
        {
            assert (mb[i].block);
            if (!*firstpos)
                *firstpos = mb[i].block;
            if (is->method->debug > 2)
                logf (LOG_LOG, "isc: skip ptr=%d size=%d %d %d",
                     i, ssize, cat, mb[i].block);
            ++(is->files[cat].no_skip_writes);
            continue;
        }
        /* write block */

        if (!*firstpos)
        {
            *firstpos = mb[i].block;
            src = r_buf + mb[i].offset - ISAMC_BLOCK_OFFSET_1;
            ssize += ISAMC_BLOCK_OFFSET_1;

            memcpy (src+sizeof(int)+sizeof(ssize), numkeys,
                    sizeof(*numkeys));
            if (is->method->debug > 2)
                logf (LOG_LOG, "isc: flush ptr=%d numk=%d size=%d nextpos=%d",
                     i, *numkeys, (int) ssize, mb[i+1].block);
        }
        else
        {
            src = r_buf + mb[i].offset - ISAMC_BLOCK_OFFSET_N;
            ssize += ISAMC_BLOCK_OFFSET_N;
            if (is->method->debug > 2)
                logf (LOG_LOG, "isc: flush ptr=%d size=%d nextpos=%d",
                     i, (int) ssize, mb[i+1].block);
        }
        memcpy (src, &mb[i+1].block, sizeof(int));
        memcpy (src+sizeof(int), &ssize, sizeof(ssize));
        isc_write_block (is, cat, mb[i].block, src);
    }
}

static int get_border (ISAMC is, struct isc_merge_block *mb, int ptr,
                       int cat, int firstpos)
{
   /* Border set to initial fill or block size depending on
      whether we are creating a new one or updating and old one.
    */
    
    int fill = mb[ptr].block ? is->method->filecat[cat].bsize :
                               is->method->filecat[cat].ifill;
    int off = (ptr||firstpos) ? ISAMC_BLOCK_OFFSET_N : ISAMC_BLOCK_OFFSET_1;
    
    assert (ptr < 199);

    return mb[ptr].offset + fill - off;
}

ISAMC_P isc_merge (ISAMC is, ISAMC_P ipos, ISAMC_I *data)
{

    char i_item[128], *i_item_ptr;
    int i_more, i_mode, i;

    ISAMC_PP pp; 
    char f_item[128], *f_item_ptr;
    int f_more;
    int last_dirty = 0;
    int debug = is->method->debug;
 
    struct isc_merge_block mb[200];

    int firstpos = 0;
    int cat = 0;
    char r_item_buf[128]; /* temporary result output */
    char *r_buf;          /* block with resulting data */
    int r_offset = 0;     /* current offset in r_buf */
    int ptr = 0;          /* pointer */
    void *r_clientData;   /* encode client data */
    int border;
    int numKeys = 0;

    r_clientData = (*is->method->code_start)(ISAMC_ENCODE);
    r_buf = is->merge_buf + 128;

    pp = isc_pp_open (is, ipos);
    /* read first item from file. make sure f_more indicates no boundary */
    f_item_ptr = f_item;
    f_more = isc_read_item (pp, &f_item_ptr);
    if (f_more > 0)
        f_more = 1;
    cat = pp->cat;

    if (debug > 1)
        logf (LOG_LOG, "isc: isc_merge begin %d %d", cat, pp->pos);

    /* read first item from i */
    i_item_ptr = i_item;
    i_more = (*data->read_item)(data->clientData, &i_item_ptr, &i_mode);

    mb[ptr].block = pp->pos;     /* is zero if no block on disk */
    mb[ptr].dirty = 0;
    mb[ptr].offset = 0;

    border = get_border (is, mb, ptr, cat, firstpos);
    while (i_more || f_more)
    {
        char *r_item = r_item_buf;
        int cmp;

        if (f_more > 1)
        {
            /* block to block boundary in the original file. */
            f_more = 1;
            if (cat == pp->cat) 
            {
                /* the resulting output is of the same category as the
                   the original 
		*/
                if (r_offset <= mb[ptr].offset +is->method->filecat[cat].mfill)
                {
                    /* the resulting output block is too small/empty. Delete
                       the original (if any)
		    */
                    if (debug > 3)
                        logf (LOG_LOG, "isc: release A");
                    if (mb[ptr].block)
                        isc_release_block (is, pp->cat, mb[ptr].block);
                    mb[ptr].block = pp->pos;
		    if (!mb[ptr].dirty)
			mb[ptr].dirty = 1;
                    if (ptr > 0)
                        mb[ptr-1].dirty = 1;
                }
                else
                {
                    /* indicate new boundary based on the original file */
                    mb[++ptr].block = pp->pos;
                    mb[ptr].dirty = last_dirty;
                    mb[ptr].offset = r_offset;
                    if (debug > 3)
                        logf (LOG_LOG, "isc: bound ptr=%d,offset=%d",
                            ptr, r_offset);
                    if (cat==is->max_cat && ptr >= is->method->max_blocks_mem)
                    {
                        /* We are dealing with block(s) of max size. Block(s)
                           except 1 will be flushed.
                         */
                        if (debug > 2)
                            logf (LOG_LOG, "isc: flush A %d sections", ptr);
                        flush_blocks (is, mb, ptr-1, r_buf, &firstpos, cat,
                                      0, &pp->numKeys);
                        mb[0].block = mb[ptr-1].block;
                        mb[0].dirty = mb[ptr-1].dirty;
                        memcpy (r_buf, r_buf + mb[ptr-1].offset,
                                mb[ptr].offset - mb[ptr-1].offset);
                        mb[0].offset = 0;

                        mb[1].block = mb[ptr].block;
                        mb[1].dirty = mb[ptr].dirty;
                        mb[1].offset = mb[ptr].offset - mb[ptr-1].offset;
                        ptr = 1;
                        r_offset = mb[ptr].offset;
                    }
                }
            }
            border = get_border (is, mb, ptr, cat, firstpos);
        }
	last_dirty = 0;
        if (!f_more)
            cmp = -1;
        else if (!i_more)
            cmp = 1;
        else
            cmp = (*is->method->compare_item)(i_item, f_item);
        if (cmp == 0)                   /* insert i=f */
        {
            if (!i_mode)   /* delete item? */
            {
                /* move i */
                i_item_ptr = i_item;
                i_more = (*data->read_item)(data->clientData, &i_item_ptr,