Re: pgsql: Support reloptions of enum type

Hi Alvaro,

On Wed, Sep 25, 2019 at 06:59:24PM +0000, Alvaro Herrera wrote:
> Support reloptions of enum type
>
> All our current in core relation options of type string (not many,
> admittedly) behave in reality like enums. But after seeing an
> implementation for enum reloptions, it's clear that strings are messier,
> so introduce the new reloption type. Switch all string options to be
> enums instead.
>
> Fortunately we have a recently introduced test module for reloptions, so
> we don't lose coverage of string reloptions, which may still be used by
> third-party modules.
>
> Authors: Nikolay Shaplov, Álvaro Herrera
> Reviewed-by: Nikita Glukhov, Aleksandr Parfenov
> Discussion: https://postgr.es/m/43332102.S2V5pIjXRx@x200m

This part from commit 773df88 is incorrect:
@@ -661,6 +700,13 @@ allocate_reloption(bits32 kinds, int type, const
char *name, const char *desc,
newoption->type = type;
newoption->lockmode = lockmode;

+ /*
+ * Set the default lock mode for this option. There is no actual way
+ * for a module to enforce it when declaring a custom relation option,
+ * so just use the highest level, which is safe for all cases.
+ */
+ newoption->lockmode = AccessExclusiveLock;

Any caller of allocate_reloption() passes down its own lockmode
definition, hence you are removing the pluggability of the API. I
think that you just got trapped by an incorrect rebase.

Do you mind if I apply the attached to fix the issue? Or perhaps you
would prefer fixing the issue yourself? I noted some inconsistencies
with the rest while on it (please see attached).
--
Michael