> I would also say that this issue is related:
>
> http://framework.zend.com/issues/browse/ZF-870
>
It is, an implementation of a transaction-nesting-counter could be
used in a resolution of this issue.
I have updated JIRA to reflect this.
> Also, can you explain what parts would be a BC break?
If the adapters, that hide the fact that the underlying engines does
not support nested transactions, were updated to throw an exception
when attempting to nest transactions, code that depends on being able
to nest transactions would break.
I do not know if this would be classified as a real BC break. The api
will not be changed. But the behavior of some adaptors, when nesting
transactions, would be changed.
Currently if you want to write code that will work across all adapters
you would have to be prepared for exceptions, because some adapters
(PDO) will throw them. What I suggest would standardize the behavior
so that all adapters throw exceptions when their underlying engine do
not support it.
Would this be classified as a BC break?
/Jacob
>
> -ralph
>
> oetting wrote:
>> I want direct some attention to an issue I have reported.
>> http://framework.zend.com/issues/browse/ZF-8063
>> The issue needs some discussion before any action can be taken. I
>> guess I
>> should have posted on the list before creating the issue, but
>> better late
>> than never :)
--
View this message in context: http://old.nabble.com/Nested-transactions-tp26284231p26289093.html
Sent from the Zend DB mailing list archive at Nabble.com.
没有评论:
发表评论