Re: [Firebird-devel] Reference counters in API

[Adriano d. S. F. <adr...@gm...>]

On 30-03-2011 07:23, Alex Peshkoff wrote:
>  On 03/29/11 18:58, Adriano dos Santos Fernandes wrote:
>> On 29-03-2011 10:45, Alex Peshkoff wrote:
>>>  We have too many problems in single thread. I try to divide it to
>>> smaller parts.
>>> The first one - here we at least all seem to talk about same thing:-) -
>>> is to use reference counters or not.
>>>
>>> On 03/28/11 22:45, Adriano dos Santos Fernandes wrote:
>>>> User should *not* request a detachment and then request actions on the
>>>> objects. This *is* user application problem, no mater it being single or
>>>> multi thread.
>>> In that case our new API is regression compared with old ISC API with
>>> handles. Pre FB3 an attempt to request actions on the objects after
>>> detachment cause correct reply from client library - invalid handle.
>>> Without reference count support we will get segfault in FB3. I treat
>>> this as serious regression.
>>>
>> You're trying to make two different worlds to look similar.
>>
>> Take another look at it.
>>
>> Thread 1 detachs and makes it handle invalid.
>>
>> Thread 2 allocs 0xFFFFFFFF handles.
>>
>> Thread 3 is scheduled to run and use the old invalid, but now valid handle.
>>
>> Very pessimistic and "impossible" to happen, but shows that application
>> doing this with handles are still incorrect.
> 
> As you understand yourself, this is not realistic example.
> Cases when thread 3 is trying to work with old invalid handle and gets
> appropriate error message is reality.
>> You also didn't replied yet to me what will happen when you detach a
>> parent without free children: leaks or will regret your refcounters?
> 
> Sorry, may be I've missed that question.
> The answer is simple - objects will exist until refcounters > 0, but
> will be marked in a way, making them return 'object is invalid' error
> for any request.
> 
And not everyone is using C++ and RAII. This will create a lot of leaks.

People would like to detach a parent and not care about statements
created but unfreed for this attachment.

And for who is using C++ and RAII, they don't need addRef/release in our
objects to use that.

> 
>>> As one of users replied: Object-based interfaces should use the refcount.
>> Against him is almost entire world who built API without refcount.
> 
> All the world is ambiguous here. See for example
> http://api.openoffice.org/docs/common/ref/com/sun/star/uno/XInterface.html
> BTW, all people who care about thread safety build object-oriented API
> with reference counters.
> 
Well, if you want to say that someone is correct just because chose the
COM way, then we would need to say we're incorrect because we don't use
queryInterface...


Adriano