Taylor Blau <me@xxxxxxxxxxxx> writes:
> callback, something like:
>
> struct option {
> /* ... */
> union {
> void *value;
> int *value_int;
> /* etc ... */
> } u;
> enum option_type t;
> };
>
> where option_type has some value corresponding to "void *", another for
> "int *", and so on.
Yup, that does cross my mind, even though I would have used
union {
void *void_ptr;
int *int_ptr;
} value;
or something without a rather meaningless 'u'.
> Alternatively, perhaps you are thinking that we'd use both the value
> pointer and the value_int pointer to point at potentially different
> values in the same callback. I don't have strong feelings about it, but
> I'd just as soon encourage us to shy away from that approach, since
> assigning a single callback parameter to each function seems more
> organized.
We have seen (with Peff's "-Wunused" work) that there are small
number of cases that it would be handy for a callback to be told the
locations of multiple external variables, but I do not think it
would be a good solution to that problem to have "void *value" and
"int value_int" next to each other and allow them to coexist, as it
would work only when these multiple variables happen to be of the
right types.
