mfile.c

harvest是一个下载html网页得机器人

/* $Id: mfile.c,v 1.52 2003/04/05 12:32:34 adam Exp $
   Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003
   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.
*/




 /*
  * TODO: The size estimates in init may not be accurate due to
  * only partially written final blocks.
  */

#include <sys/types.h>
#include <fcntl.h>
#ifdef WIN32
#include <io.h>
#else
#include <unistd.h>
#endif
#include <direntz.h>
#include <string.h>
#include <stdlib.h>
#include <stdio.h>
#include <assert.h>
#include <errno.h>

#include <zebra-lock.h>
#include <zebrautl.h>
#include <mfile.h>

static int scan_areadef(MFile_area ma, const char *ad, const char *base)
{
    /*
     * If no definition is given, use current directory, unlimited.
     */
    char dirname[FILENAME_MAX+1]; 
    mf_dir **dp = &ma->dirs, *dir = *dp;

    if (!ad)
        ad = ".:-1b";
    for (;;)
    {
        const char *ad0 = ad;
        int i = 0, fact = 1, multi;
	mfile_off_t size = 0;

        while (*ad == ' ' || *ad == '\t')
            ad++;
        if (!*ad)
            break;
        if (!yaz_is_abspath(ad) && base)
        {
            strcpy (dirname, base);
            i = strlen(dirname);
            dirname[i++] = '/';
        }
        while (*ad)
        {
	    if (*ad == ':' && strchr ("+-0123456789", ad[1]))
		break;
            if (i < FILENAME_MAX)
                dirname[i++] = *ad;
            ad++;
        }
        dirname[i] = '\0';
        if (*ad++ != ':')
        {
	    logf (LOG_WARN, "Missing colon after path: %s", ad0);
            return -1;
        }
        if (i == 0)
        {
	    logf (LOG_WARN, "Empty path: %s", ad0);
            return -1;
        }
        while (*ad == ' ' || *ad == '\t')
            ad++;
        if (*ad == '-')
        {
            fact = -1;
            ad++;
        }
        else if (*ad == '+')
            ad++;
        size = 0;
        if (*ad < '0' || *ad > '9')
        {
	    logf (LOG_FATAL, "Missing size after path: %s", ad0);
            return -1;
        }
        size = 0;
        while (*ad >= '0' && *ad <= '9')
            size = size*10 + (*ad++ - '0');
        switch (*ad)
	{
	case 'B': case 'b': multi = 1; break;
	case 'K': case 'k': multi = 1024; break;
	case 'M': case 'm': multi = 1048576; break;
	case 'G': case 'g': multi = 1073741824; break;
        case '\0':
            logf (LOG_FATAL, "Missing unit: %s", ad0);
            return -1;
        default:
            logf (LOG_FATAL, "Illegal unit: %c in %s", *ad, ad0);
            return -1;
	}
        ad++;
	*dp = dir = (mf_dir *) xmalloc(sizeof(mf_dir));
	dir->next = 0;
	strcpy(dir->name, dirname);
	dir->max_bytes = dir->avail_bytes = fact * size * multi;
	dp = &dir->next;
    }
    return 0;
}

static int file_position(MFile mf, int pos, int offset)
{
    int off = 0, c = mf->cur_file, ps;

    if ((c > 0 && pos <= mf->files[c-1].top) ||
	(c < mf->no_files -1 && pos > mf->files[c].top))
    {
	c = 0;
	while (c + 1 < mf->no_files && mf->files[c].top < pos)
	{
	    off += mf->files[c].blocks;
	    c++;
	}
	assert(c < mf->no_files);
    }
    else
    	off = c ? (mf->files[c-1].top + 1) : 0;
    if (mf->files[c].fd < 0)
    {
        if ((mf->files[c].fd = open(mf->files[c].path,
	                            mf->wr ?
                                        (O_BINARY|O_RDWR|O_CREAT) :
                                        (O_BINARY|O_RDONLY), 0666)) < 0)
        {
            if (!mf->wr && errno == ENOENT && off == 0)
                return -2;
    	    logf (LOG_WARN|LOG_ERRNO, "Failed to open %s", mf->files[c].path);
    	     return -1;
        }
    }
    ps = pos - off;
    if (mfile_seek(mf->files[c].fd, ps * (mfile_off_t) mf->blocksize + offset,
    	SEEK_SET) < 0)
    {
    	logf (LOG_WARN|LOG_ERRNO, "Failed to seek in %s", mf->files[c].path);
        logf(LOG_WARN, "pos=%d off=%d blocksize=%d offset=%d",
                       pos, off, mf->blocksize, offset);
    	return -1;
    }
    mf->cur_file = c;
    return ps;
}

static int cmp_part_file(const void *p1, const void *p2)
{
    return ((part_file *)p1)->number - ((part_file *)p2)->number;
}

/*
 * Create a new area, cotaining metafiles in directories.
 * Find the part-files in each directory, and inventory the existing metafiles.
 */
MFile_area mf_init(const char *name, const char *spec, const char *base)
{
    MFile_area ma = (MFile_area) xmalloc(sizeof(*ma));
    mf_dir *dirp;
    meta_file *meta_f;
    part_file *part_f = 0;
    DIR *dd;
    struct dirent *dent;
    int fd, number;
    char metaname[FILENAME_MAX+1], tmpnam[FILENAME_MAX+1];
    
    logf (LOG_DEBUG, "mf_init(%s)", name);
    strcpy(ma->name, name);
    ma->mfiles = 0;
    ma->dirs = 0;
    if (scan_areadef(ma, spec, base) < 0)
    {
    	logf (LOG_WARN, "Failed to access description of '%s'", name);
    	return 0;
    }
    /* look at each directory */
    for (dirp = ma->dirs; dirp; dirp = dirp->next)
    {
    	if (!(dd = opendir(dirp->name)))
    	{
    	    logf (LOG_WARN|LOG_ERRNO, "Failed to open directory %s",
                                     dirp->name);
    	    return 0;
	}
	/* look at each file */
	while ((dent = readdir(dd)))
	{
	    int len = strlen(dent->d_name);
	    const char *cp = strrchr (dent->d_name, '-');
	    if (strchr (".-", *dent->d_name))
		continue;
	    if (len < 5 || !cp || strcmp (dent->d_name + len - 3, ".mf"))
		continue;
	    number = atoi(cp+1);
	    memcpy (metaname, dent->d_name, cp - dent->d_name);
	    metaname[ cp - dent->d_name] = '\0';

	    for (meta_f = ma->mfiles; meta_f; meta_f = meta_f->next)
	    {
	    	/* known metafile */
		if (!strcmp(meta_f->name, metaname))
	    	{
		    part_f = &meta_f->files[meta_f->no_files++];
		    break;
		}
	    }
	    /* new metafile */
	    if (!meta_f)
	    {
	    	meta_f = (meta_file *) xmalloc(sizeof(*meta_f));
        	zebra_mutex_init (&meta_f->mutex);
	    	meta_f->ma = ma;
	    	meta_f->next = ma->mfiles;
	    	meta_f->open = 0;
	    	meta_f->cur_file = -1;
                meta_f->unlink_flag = 0;
	    	ma->mfiles = meta_f;
	    	strcpy(meta_f->name, metaname);
	    	part_f = &meta_f->files[0];
	    	meta_f->no_files = 1;
	    }
	    part_f->number = number;
	    part_f->dir = dirp;
	    part_f->fd = -1;
	    sprintf(tmpnam, "%s/%s", dirp->name, dent->d_name);
	    part_f->path = xstrdup(tmpnam);
	    /* get size */
	    if ((fd = open(part_f->path, O_BINARY|O_RDONLY)) < 0)
	    {
	    	logf (LOG_FATAL|LOG_ERRNO, "Failed to access %s",
                      dent->d_name);
	    	return 0;
	    }
	    if ((part_f->bytes = mfile_seek(fd, 0, SEEK_END)) < 0)
	    {
	    	logf (LOG_FATAL|LOG_ERRNO, "Failed to seek in %s",
                      dent->d_name);
	    	return 0;
	    }
#ifndef WIN32
	    fsync(fd);
#endif
	    close(fd);
	    if (dirp->max_bytes >= 0)
		dirp->avail_bytes -= part_f->bytes;
	}
	closedir(dd);
    }
    for (meta_f = ma->mfiles; meta_f; meta_f = meta_f->next)
    {
    	logf (LOG_DEBUG, "mf_init: %s consists of %d part(s)", meta_f->name,
              meta_f->no_files);
    	qsort(meta_f->files, meta_f->no_files, sizeof(part_file),
              cmp_part_file);
    }
    return ma;
}

void mf_destroy(MFile_area ma)
{
    mf_dir *dp;
    meta_file *meta_f;

    if (!ma)
	return;
    dp = ma->dirs;
    while (dp)
    {
	mf_dir *d = dp;
	dp = dp->next;
	xfree (d);
    }
    meta_f = ma->mfiles;
    while (meta_f)
    {
	int i;