Hi,
On 2024-05-15 13:45:30 -0400, Tom Lane wrote:
> There is one advantage over my suggestion of changing PG_MODULE_MAGIC:
> if we tell people to write
>
> PG_MODULE_MAGIC;
> #undef TEXTDOMAIN
> #define TEXTDOMAIN PG_TEXTDOMAIN("hstore")
>
> then that's 100% backwards compatible and they don't need any
> version-testing ifdef's.
>
> I still think that the kind of infrastructure Andres proposes
> is way overkill compared to the value, plus it's almost certainly
> going to have a bunch of platform-specific problems to solve.
Maybe I missing something here. Even adding those two lines to the extensions
in core and contrib is going to end up being more lines than what I proposed?
What portability issues do you forsee? We already look up the same symbol in
all the shared libraries ("Pg_magic_func"), so we know that we can deal with
duplicate function names. Are you thinking that somehow we'd end up with
symbol interposition or something?
Greetings,
Andres Freund