Re: proposal: enable new error fields in plpgsql (9.4)

2013/2/1 Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>:
> 2013/2/1 Peter Eisentraut <peter_e(at)gmx(dot)net>:
>> On 2/1/13 8:00 AM, Pavel Stehule wrote:
>>> 2013/2/1 Marko Tiikkaja <pgmail(at)joh(dot)to>:
>>>> On 2/1/13 1:47 PM, Pavel Stehule wrote:
>>>>>
>>>>> now a most "hard" work is done and I would to enable access to new
>>>>> error fields from plpgsql.
>>>>
>>>>
>>>> Is there a compelling reason why we wouldn't provide these already in 9.3?
>>>
>>> a time for assign to last commitfest is out.
>>>
>>> this patch is relative simple and really close to enhanced error
>>> fields feature - but depends if some from commiters will have a time
>>> for commit to 9.3 - so I am expecting primary target 9.4, but I am not
>>> be angry if it will be commited early.
>>
>> If we don't have access to those fields on PL/pgSQL, what was the point
>> of the patch to begin with? Surely, accessing them from C wasn't the
>> main use case?
>>
>
> These fields are available for application developers now. But is a
> true, so without this patch, GET STACKED DIAGNOSTICS statement will
> not be fully consistent, because some fields are accessible and others
> not

there is one stronger argument for commit this patch now. With this
patch, we are able to wrote regression tests for new fields via
plpgsql.

Regards

Pavel

>
> Pavel