chm_lib.c

又一开源的解压缩chm格式文件的代码

                /* decompress the previous block */
#ifdef CHM_DEBUG
                fprintf(stderr, "Decompressing block #%4d (EXTRA)\n", curBlockIdx);
#endif
                if (!_chm_get_cmpblock_bounds(h, curBlockIdx, &cmpStart, &cmpLen) ||
                    _chm_fetch_bytes(h, cbuffer, cmpStart, cmpLen) != cmpLen      ||
                    LZXdecompress(h->lzx_state, cbuffer, lbuffer, (int)cmpLen,
                                  (int)h->reset_table.block_len) != DECR_OK)
                {
#ifdef CHM_DEBUG
                    fprintf(stderr, "   (DECOMPRESS FAILED!)\n");
#endif
                    FREEBUF(cbuffer);
                    return (Int64)0;
                }

                h->lzx_last_block = (int)curBlockIdx;
            }
        }
    }
    else
    {
        if ((block % h->reset_blkcount) == 0)
        {
#ifdef CHM_DEBUG
            fprintf(stderr, "***RESET (2)***\n");
#endif
            LZXreset(h->lzx_state);
        }
    }

    /* allocate slot in cache */
    indexSlot = (int)(block % h->cache_num_blocks);
    h->cache_block_indices[indexSlot] = block;
    if (! h->cache_blocks[indexSlot])
        h->cache_blocks[indexSlot] = (UChar *)malloc(
                                          ((unsigned int)h->reset_table.block_len));
    lbuffer = h->cache_blocks[indexSlot];
    *ubuffer = lbuffer;

    /* decompress the block we actually want */
#ifdef CHM_DEBUG
    fprintf(stderr, "Decompressing block #%4d (REAL )\n", block);
#endif
    if (! _chm_get_cmpblock_bounds(h, block, &cmpStart, &cmpLen)          ||
        _chm_fetch_bytes(h, cbuffer, cmpStart, cmpLen) != cmpLen          ||
        LZXdecompress(h->lzx_state, cbuffer, lbuffer, (int)cmpLen,
                      (int)h->reset_table.block_len) != DECR_OK)
    {
#ifdef CHM_DEBUG
        fprintf(stderr, "   (DECOMPRESS FAILED!)\n");
#endif
        FREEBUF(cbuffer);
        return (Int64)0;
    }
    h->lzx_last_block = (int)block;

    /* XXX: modify LZX routines to return the length of the data they
     * decompressed and return that instead, for an extra sanity check.
     */
    FREEBUF(cbuffer);
    return h->reset_table.block_len;
}

/* grab a region from a compressed block */
static Int64 _chm_decompress_region(struct chmFile *h,
                                    UChar *buf,
                                    UInt64 start,
                                    Int64 len)
{
    UInt64 nBlock, nOffset;
    UInt64 nLen;
    UInt64 gotLen;
    UChar *ubuffer;

        if (len <= 0)
                return (Int64)0;

    /* figure out what we need to read */
    nBlock = start / h->reset_table.block_len;
    nOffset = start % h->reset_table.block_len;
    nLen = len;
    if (nLen > (h->reset_table.block_len - nOffset))
        nLen = h->reset_table.block_len - nOffset;

    /* if block is cached, return data from it. */
    CHM_ACQUIRE_LOCK(h->lzx_mutex);
    CHM_ACQUIRE_LOCK(h->cache_mutex);
    if (h->cache_block_indices[nBlock % h->cache_num_blocks] == nBlock    &&
        h->cache_blocks[nBlock % h->cache_num_blocks] != NULL)
    {
        memcpy(buf,
               h->cache_blocks[nBlock % h->cache_num_blocks] + nOffset,
               (unsigned int)nLen);
        CHM_RELEASE_LOCK(h->cache_mutex);
        CHM_RELEASE_LOCK(h->lzx_mutex);
        return nLen;
    }
    CHM_RELEASE_LOCK(h->cache_mutex);

    /* data request not satisfied, so... start up the decompressor machine */
    if (! h->lzx_state)
    {
        int window_size = ffs(h->window_size) - 1;
        h->lzx_last_block = -1;
        h->lzx_state = LZXinit(window_size);
    }

    /* decompress some data */
    gotLen = _chm_decompress_block(h, nBlock, &ubuffer);
    if (gotLen < nLen)
        nLen = gotLen;
    memcpy(buf, ubuffer+nOffset, (unsigned int)nLen);
    CHM_RELEASE_LOCK(h->lzx_mutex);
    return nLen;
}

/* retrieve (part of) an object */
LONGINT64 chm_retrieve_object(struct chmFile *h,
                               struct chmUnitInfo *ui,
                               unsigned char *buf,
                               LONGUINT64 addr,
                               LONGINT64 len)
{
    /* must be valid file handle */
    if (h == NULL)
        return (Int64)0;

    /* starting address must be in correct range */
    if (addr < 0  ||  addr >= ui->length)
        return (Int64)0;

    /* clip length */
    if (addr + len > ui->length)
        len = ui->length - addr;

    /* if the file is uncompressed, it's simple */
    if (ui->space == CHM_UNCOMPRESSED)
    {
        /* read data */
        return _chm_fetch_bytes(h,
                                buf,
                                (UInt64)h->data_offset + (UInt64)ui->start + (UInt64)addr,
                                len);
    }

    /* else if the file is compressed, it's a little trickier */
    else /* ui->space == CHM_COMPRESSED */
    {
        Int64 swath=0, total=0;

        /* if compression is not enabled for this file... */
        if (! h->compression_enabled)
            return total;

        do {

            /* swill another mouthful */
            swath = _chm_decompress_region(h, buf, ui->start + addr, len);

            /* if we didn't get any... */
            if (swath == 0)
                return total;

            /* update stats */
            total += swath;
            len -= swath;
            addr += swath;
            buf += swath;

        } while (len != 0);

        return total;
    }
}

/* enumerate the objects in the .chm archive */
int chm_enumerate(struct chmFile *h,
                  int what,
                  CHM_ENUMERATOR e,
                  void *context)
{
    Int32 curPage;

    /* buffer to hold whatever page we're looking at */
#ifdef WIN32
#ifdef _WIN32_WCE
    /* RWE 6/12/2003 */
    UChar *page_buf = malloc((unsigned int)h->block_len);
#else
    UChar *page_buf = alloca((unsigned int)h->block_len);
#endif
#else
    UChar page_buf[h->block_len];
#endif
    struct chmPmglHeader header;
    UChar *end;
    UChar *cur;
    unsigned long lenRemain;
    UInt64 ui_path_len;

    /* the current ui */
    struct chmUnitInfo ui;
    int flag;

    /* starting page */
    curPage = h->index_head;

    /* until we have either returned or given up */
    while (curPage != -1)
    {

        /* try to fetch the index page */
        if (_chm_fetch_bytes(h,
                             page_buf,
                             (UInt64)h->dir_offset + (UInt64)curPage*h->block_len,
                             h->block_len) != h->block_len)
        {
            FREEBUF(page_buf);
            return 0;
        }

        /* figure out start and end for this page */
        cur = page_buf;
        lenRemain = _CHM_PMGL_LEN;
        if (! _unmarshal_pmgl_header(&cur, &lenRemain, &header))
        {
            FREEBUF(page_buf);
            return 0;
        }
        end = page_buf + h->block_len - (header.free_space);

        /* loop over this page */
        while (cur < end)
        {
            if (! _chm_parse_PMGL_entry(&cur, &ui))
            {
                FREEBUF(page_buf);
                return 0;
            }

            /* get the length of the path */
            ui_path_len = strlen(ui.path)-1;

            /* check for DIRS */
            if (ui.path[ui_path_len] == '/'  &&  !(what & CHM_ENUMERATE_DIRS))
                continue;

            /* check for FILES */
            if (ui.path[ui_path_len] != '/'  &&  !(what & CHM_ENUMERATE_FILES))
                continue;

            /* check for NORMAL vs. META */
            if (ui.path[0] == '/')
            {

                /* check for NORMAL vs. SPECIAL */
                if (ui.path[1] == '#'  ||  ui.path[1] == '$')
                    flag = CHM_ENUMERATE_SPECIAL;
                else
                    flag = CHM_ENUMERATE_NORMAL;
            }
            else
                flag = CHM_ENUMERATE_META;
            if (! (what & flag))
                continue;

            /* call the enumerator */
            {
                int status = (*e)(h, &ui, context);
                switch (status)
                {
                    case CHM_ENUMERATOR_FAILURE:
                        FREEBUF(page_buf);
                        return 0;
                    case CHM_ENUMERATOR_CONTINUE:
                        break;
                    case CHM_ENUMERATOR_SUCCESS:
                        FREEBUF(page_buf);
                        return 1;
                    default:
                        break;
                }
            }
        }

        /* advance to next page */
        curPage = header.block_next;
    }

    FREEBUF(page_buf);
    return 1;
}

int chm_enumerate_dir(struct chmFile *h,
                      const char *prefix,
                      int what,
                      CHM_ENUMERATOR e,
                      void *context)
{
    /*
     * XXX: do this efficiently (i.e. using the tree index)
     */

    Int32 curPage;

    /* buffer to hold whatever page we're looking at */
#ifdef WIN32
#ifdef _WIN32_WCE
    /* RWE 6/12/2003 */
    UChar *page_buf = malloc((unsigned int)h->block_len);
#else
    UChar *page_buf = alloca((unsigned int)h->block_len);
#endif
#else
    UChar page_buf[h->block_len];
#endif
    struct chmPmglHeader header;
    UChar *end;
    UChar *cur;
    unsigned long lenRemain;

    /* set to 1 once we've started */
    int it_has_begun=0;

    /* the current ui */
    struct chmUnitInfo ui;
    int flag;
    UInt64 ui_path_len;

    /* the length of the prefix */
    char prefixRectified[CHM_MAX_PATHLEN+1];
    int prefixLen;
    char lastPath[CHM_MAX_PATHLEN];
    int lastPathLen;

    /* starting page */
    curPage = h->index_head;

    /* initialize pathname state */
    strncpy(prefixRectified, prefix, CHM_MAX_PATHLEN);
    prefixLen = strlen(prefixRectified);
    if (prefixLen != 0)
    {
        if (prefixRectified[prefixLen-1] != '/')
        {
            prefixRectified[prefixLen] = '/';
            prefixRectified[prefixLen+1] = '\0';
            ++prefixLen;
        }
    }
    lastPath[0] = '\0';
    lastPathLen = -1;

    /* until we have either returned or given up */
    while (curPage != -1)
    {

        /* try to fetch the index page */
        if (_chm_fetch_bytes(h,
                             page_buf,
                             (UInt64)h->dir_offset + (UInt64)curPage*h->block_len,
                             h->block_len) != h->block_len)
        {
            FREEBUF(page_buf);
            return 0;
        }

        /* figure out start and end for this page */
        cur = page_buf;
        lenRemain = _CHM_PMGL_LEN;
        if (! _unmarshal_pmgl_header(&cur, &lenRemain, &header))
        {
            FREEBUF(page_buf);
            return 0;
        }
        end = page_buf + h->block_len - (header.free_space);

        /* loop over this page */
        while (cur < end)
        {
            if (! _chm_parse_PMGL_entry(&cur, &ui))
            {
                FREEBUF(page_buf);
                return 0;
            }

            /* check if we should start */
            if (! it_has_begun)
            {
                if (ui.length == 0  &&  strncasecmp(ui.path, prefixRectified, prefixLen) == 0)
                    it_has_begun = 1;
                else
                    continue;

                if (ui.path[prefixLen] == '\0')
                    continue;
            }

            /* check if we should stop */
            else
            {
                if (strncasecmp(ui.path, prefixRectified, prefixLen) != 0)
                {
                    FREEBUF(page_buf);
                    return 1;
                }
            }

            /* check if we should include this path */
            if (lastPathLen != -1)
            {
                if (strncasecmp(ui.path, lastPath, lastPathLen) == 0)
                    continue;
            }
            strcpy(lastPath, ui.path);
            lastPathLen = strlen(lastPath);

            /* get the length of the path */
            ui_path_len = strlen(ui.path)-1;

            /* check for DIRS */
            if (ui.path[ui_path_len] == '/'  &&  !(what & CHM_ENUMERATE_DIRS))
                continue;

            /* check for FILES */
            if (ui.path[ui_path_len] != '/'  &&  !(what & CHM_ENUMERATE_FILES))
                continue;

            /* check for NORMAL vs. META */
            if (ui.path[0] == '/')
            {

                /* check for NORMAL vs. SPECIAL */
                if (ui.path[1] == '#'  ||  ui.path[1] == '$')
                    flag = CHM_ENUMERATE_SPECIAL;
                else
                    flag = CHM_ENUMERATE_NORMAL;
            }
            else
                flag = CHM_ENUMERATE_META;
            if (! (what & flag))
                continue;

            /* call the enumerator */
            {
                int status = (*e)(h, &ui, context);
                switch (status)
                {
                    case CHM_ENUMERATOR_FAILURE:
                        FREEBUF(page_buf);
                        return 0;
                    case CHM_ENUMERATOR_CONTINUE:
                        break;
                    case CHM_ENUMERATOR_SUCCESS:
                        FREEBUF(page_buf);
                        return 1;
                    default:
                        break;
                }
            }
        }

        /* advance to next page */
        curPage = header.block_next;
    }

    FREEBUF(page_buf);
    return 1;
}