Vlad Horsun wrote:

>>>> And Vulcan's config stuff can't prevent such problem too - we can open the
>>>> same database via Firebird and via Vulcan simultaneously
>>>>
>>>>
>>> And i did it.
>>>
>>> Are nobody worried about it ?
>>>
>>>
>>>
>> Hey, did it work?
>>
>
> At least i can do SELECT's in both (Firebird's and Vulcan's) sessions.
> Not tried to corrupt database writting something :)
>
>
>> Seriously, detecting the case where two engines open the same file with
>> different lock files is well worth solving. I think a solution where
>> the first database file opener writes the same random number into the
>> database header and lock block header, then downgrades the file system
>> lock from exclusive to shared. Anyone who couldn't get an exclusive
>> lock on the file would check the header page against the lock table.
>>
>
> Yes, this will work. But only for new engines which will make such checks.
> I'm afraid we need to change ODS number to make it work. Currently Vulcan
> support ODS 11. If we implement this check in Firebird 2.1 (as it is planned
> to support multy instance and already have ODS 11.1) then i suggest to
> make Vulcan support ODS starting from 11.1 (if we plan to make public
> release of Vulcan before merge)
>
>

Bumping the minor version number is effective and painless. Why not?