> > I notice that all the functions with the same name but different args
> > are actually sql statements which SELECT the result of the function
> > call using a different (and unique) name..
> > Wouldn't this cause slowdowns? Shouldn't you be able to have a
> > different name for your function in pgsql than in the shared library,
> > without having to resort to such hacks?
> Actually, we were pretty happy when Edmund Mergl found this mechanism.
> I've thought about making changes to allow compiled code to do the same
> thing, but we've had other more important issues to work on. Send
> patches if you want something different.
Actually the problem was that SQL functions can compare args and call
the proper function, while C functions just get called without any arg
In response to
pgsql-hackers by date
|Next:||From: Simon Shapiro||Date: 1998-01-28 16:50:51|
|Subject: Re: [HACKERS] Domain Problem?|
|Previous:||From: Thomas G. Lockhart||Date: 1998-01-28 16:22:28|
|Subject: Re: [HACKERS] No: implied sort with group by|