Re: Problem with aborting entire transactions on error

On Dec 11, 2012, at 6:28, Zbigniew <zbigniew2011(at)gmail(dot)com> wrote:

> 2012/12/11, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
>
>> The complexity and performance costs that people have mentioned are other
>> good reasons not to change it; but even if the change were free on the
>> server side, history says it's not something we ought to mess with.
>
> Again: my suggestion was to add an option... with default value "keep
> present functionality".
>

We heard you the first three times ...

All of the responses given so far, including the paragraph you are quoting, are given with that in mind. You have made your point clearly and have been given direct responses.

There is some merit but not enough to convince anyone to change their general belief so now it comes down to providing a specific implementation that can be approved or rejected as belonging in core or go with the various alternatives that have been presented.

David J.