Hi,
On 2018-10-16 12:41:44 -0400, Robert Haas wrote:
> On Tue, Oct 16, 2018 at 12:26 PM Andres Freund <andres@anarazel.de> wrote:
> > /*
> > * Macro that allows to cast constness away from a variable, but doesn't
> > * allow changing the underlying type. Enforcement of the latter
> > * currently only works for gcc like compilers.
> > *
> > * Please note IT IS NOT SAFE to cast constness away if the variable will ever
> > * be modified (it would be undefined behaviour). Doing so anyway can cause
> > * compiler misoptimizations or runtime crashes (modifying readonly memory).
> > * It is only safe to use when the the variabble will not be modified, but API
> > * design or language restrictions prevent you from declaring that
> > * (e.g. because a function returns both const and non-const variables).
> > */
>
> "variabble" is a little too rich in "b"s.
variababble.
> In terms of a function that returns both const and non-const
> variables, it seems a bit sketchy that the caller would know what the
> function is doing in particular cases and make decisions based on it,
> but maybe that's just how life is.
I don't think it's necessary the callers doing so in most cases. E.g. in
the DestReceiver case, it'll be the choice of the testreceiver (say
intorel_receive modifying things for DestIntoRel), not the caller
choosing when to modify things. The caller / users of dest receivers
won't necessarily know.
I agree it's not pretty, but I don't quite see any really realistic
other approaches here.
Greetings,
Andres Freund