Re: proposal: better support for debugging of overloaded functions

Hello

2011/11/21 Robert Haas <robertmhaas(at)gmail(dot)com>:
> On Sun, Nov 20, 2011 at 6:16 AM, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> wrote:
>> Is possible to add GUC variable plpgsql.log_function_signature (maybe
>> just log_function_signature (for all PL))? I am not sure about GUC
>> name.
>>
>> When this variable is true, then CONTEXT line will contain a qualified
>> function's signature instead function name
>
> Sure, but why?  If it's possible to do that, I think we should just do
> it always.  It might be a net reduction in readability for people who
> don't use overloading but do have functions with very long names and
> lots and lots of arguments, but even if you think that's good design,
> I think the general principle that an error message should uniquely
> identify the object responsible for the error ought to take
> precedence.

I inclined so this is good solution

there is a VIP patch

patch is relative long, but almost all are changes in regress tests.
Changes in plpgsql are 5 lines

Regards

Pavel

>
> --
> Robert Haas
> EnterpriseDB: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company
>