Ah right, I nearly forgot that - overconfidence from my incomplete test.
Shall I completely remove the implementation or just comment on it and
remove it from being a default?
Sanne
2010/4/5 Emmanuel Bernard <emmanuel(a)hibernate.org>:
The problem with the exception impl is that it does not work today
because Core's Synch does swallow exceptions. See "Exceptions thrown in a tx
synchronization are eaten".
I am not sure we should put it in until this is resolved.
On 4 avr. 2010, at 18:34, Sanne Grinovero wrote:
> Emmanuel,
> I'm applying the latest patch from Amin, the patch is fine but
> thinking more on the async/sync I realized we forgot the other
> implementation, so I added an ErrorHandler implementation which will
> rethrow the exception up to the thread which committed the
> transaction, and another test for this.
>
> Could you check it, especially the configuration option key and names
> for the two error handlers?
>
> To close the issue, I'll wait for Amin to write a little documentation patch.
>
> Sanne
>
> 2010/4/3 Sanne Grinovero <sanne.grinovero(a)gmail.com>:
>> 2010/4/3 Emmanuel Bernard <emmanuel(a)hibernate.org>:
>>> Ah cool, we are getting there :)
>>> Looking at it quickly, can you use tabs, not space for indentation?
That's our coding conventions.
>>>
>>> One open question I had last time was one error handler for sync and async,
vs one for sync and one for async. What was decided? What was the reasoning?
>>>
>>
>> Initially we thought that that would be needed, but doesn't appear to
>> be the case. Latest Amin's work did remove the duality, sync and async
>> are using the same methods now.
>>