From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: cast result of copyNode() |
Date: | 2016-12-31 16:56:01 |
Message-ID: | 32425.1483203361@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> writes:
> In order to reduce the number of useless casts and make the useful casts
> more interesting, here is a patch that automatically casts the result of
> copyNode() back to the input type, if the compiler supports something
> like typeof(), which most current compilers appear to. That gets us
> some more type safety and we only need to retain the casts that actually
> do change the type.
But doesn't this result in a boatload of warnings on compilers that
don't have typeof()?
If typeof() were in C99 I might figure that is an okay tradeoff,
but it isn't. I'm not sure that this is the most useful place
to be moving our C standards compliance expectations by 20 years
in one jump.
Also, if your answer is "you shouldn't get any warnings because
copyObject is already declared to return void *", then why aren't
we just relying on that today?
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Peter Eisentraut | 2016-12-31 17:08:22 | safer node casting |
Previous Message | Fabien COELHO | 2016-12-31 16:51:20 | Re: proposal: session server side variables |