Claudio Valderrama C. wrote:

>> -----Original Message-----
>> From: Firebird-Architect@yahoogroups.com
>> [mailto:Firebird-Architect@yahoogroups.com]On Behalf Of Ann W. Harrison
>> Sent: Martes, 19 de Septiembre de 2006 18:14
>>
>> Martijn Tonies wrote:
>>
>>>> Can you think of any reason why a developer would create a trigger on
>>>> the RDB$Database table?
>>>>
>>> Well, given that triggers on system table don't "stick", no.
>>>
>>>
>> That could be fixed. It's really a gbak problem / quirk.
>>
>
> Also, user triggers on sys tables work only until everybody disconnects from
> that db.
> When you define a user trigger in a sys table, it's loaded and works.
> After all attachments to that db are closed, the server "unloads" the db.
> If you connect again, when the server loads the db, it doesn't load user
> triggers on sys tables.
>
>

True, that's that not an implementation artifact, not an architectural
characteristic. System table are built from compiled in tables rather
than loaded from metadata (chicken / egg problem). That could mechanism
could be extended to check for triggers on the core tables. Maybe take
an hour to implement. But that doesn't address the backup problem, however.

--

Jim Starkey
Netfrastructure, Inc.
978 526-1376