Re: [mdb-dev] Crash in mdb-import

[<bo...@ma...>]

Thanks, but i think there were no insert/remove when i created the=20
file. I'll double check. Thanks.

Regards,
Mikl=F3s
On Mar 16, 2005, at 2:10 PM, Sam Moffatt wrote:

> What could be causing the result to be 21 is if rows have been deleted
> from the table and rows added afterwards. 21 could be the index of th
> e last row. I might be wrong, but it would appear to make sense.
> Someone with a bit more time might be able to help,
>
> Sam
>
> On Wed, 16 Mar 2005 08:17:09 +0100, Mikl=F3s Fazekas <bo...@ma...>=20
> wrote:
>> I've tried to use mdb-import for a simple jet4 db.
>> The result was a crash in the following code:
>>
>> (The sample file can be found attached to the bug:
>> http://sourceforge.net/tracker/index.php?
>> func=3Ddetail&aid=3D1163895&group_id=3D2294&atid=3D102294
>> )
>>
>>   } else { /* is not a temp table */
>>         new_pg =3D mdb_new_data_pg(entry);
>>
>>          num_rows =3D mdb_pg_get_int16(mdb, fmt->row_count_offset);
>>          pos =3D fmt->pg_size;
>>
>>         /* copy existing rows */
>>         for (i=3D0;i<num_rows;i++) {
>>                 mdb_find_row(mdb, i, &row_start, &row_size);
>>                 pos -=3D row_size;
>>
>>                 memcpy(&new_pg[pos], &mdb->pg_buf[row_start],=20
>> row_size);
>> ^^^^^^^^^^^^^-crash for i =3D=3D 5
>>                 _mdb_put_int16(new_pg, (fmt->row_count_offset + 2) +=20=

>> (i*2), pos);
>>          }
>>   }
>>
>> Any hints?!
>> For me the num_rows seems to be too large (it's 21 while the table =
has
>> only 4 rows), but i might be completly wrong.
>> [ And setting the num_rows to 4 solved the crash but the result was
>> still a bit bogus... ]
>>
>> Regards,
>> Mikl=F3s
>>
>>