Re: R: Re: R: Re: Weird EXECUTE ... USING behaviour

2010/1/15 Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>:
> 2010/1/15 Vincenzo Romano <vincenzo(dot)romano(at)notorand(dot)it>:
>> 2010/1/15 Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>:
>>>>
>>>> 3. Then I need to add the TABLE-level CHECK condition in order to
>>>> exploit the "constraint_exclusion = on":
>>>>            execute $l2$
>>>>              alter table $l2$||ct||$l2$
>>>>                add check(  $l0$||co||$l0$ )
>>>>            $l2$ using $l0$||va||$l0$;
>>>>
>>>
>>> I am sorry, I am out. Your code isn't much readable:
>>>
>>> Minimally there isn't placeholder - some like $x
>>>
>>> It is game for the most cryptografic code :).
>>>
>>> You searching a functionality that isn't in pg now :(. Clause USING
>>> doesn't work like macros or templates in C++
>>>
>>> look on this page - maybe it could be useful for you
>>> http://wiki.postgresql.org/wiki/Sprintf
>>>
>>> Regards
>>> Pavel
>>>
>>
>> Pavel,
>> It's worse than cryptographic if you skip the starting points. I do
>> admit it's far from plain PLPGSQL, though.
>> The "co" text variable contains the $x placeholders, the "va" text
>> variable contains the expressions to be used.
>
> this is bad. Placeholders are used only when are explicit - tj they
> are in string constant. Content of variables are protected.

sorry - it isn't exactly true. depend on nesting level.

Pavel

>
> Pavel
>
>>
>> --
>> Vincenzo Romano
>> NotOrAnd Information Technologies
>> cel. +39 339 8083886  | gtalk. vincenzo(dot)romano(at)notorand(dot)it
>> fix. +39 0823 454163  | skype. notorand.it
>> fax. +39 02 700506964 | msn.   notorand.it
>> NON QVIETIS MARIBVS NAVTA PERITVS
>>
>