13.02.2012 13:45, Alex Peshkoff wrote:

>> I'm not really sure this is strictly required for v2.x. We surely must
>> add an overflow protection, here I agree.
>
> A check in shared memory grow code should do a trick.

Sure.

>> But if it would be possible to
>> have ULONG offsets instead of SLONG, it would double the limit. RedSoft
>> could have their 2500 connections safely and nobody else is affected, so
>> this sounds as a good compromise for a point release, IMHO.
>
> May be use 1,2,3 instead -1,-2,-3 for 'magics' like DUMMY_OWNER?
> IMHO real offsets can't be so small.

Interesting idea. But how should the conditions like (offset < 0) look 
now? Something like (offset < sizeof(lhb)) doesn't look elegant. Maybe 
explicitly compare with all the magic numbers?

>>> What about FB3 - may be we should start to think about lock-manager as
>>> replaceable dynamic library?
>> I'm thinking about it for a long time already :-) And yes, it's likely
>> to happen in v3.
>
> The worst is the moment when that library is going to be unloaded. All
> the rest looks quite easy to do.

We already have a refcount for the lock manager.


Dmitry

------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel

Reply via email to