Re: AllocSetContextCreate changes breake extensions

Andres Freund <andres@anarazel.de>

From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@postgresql.org, Christoph Berg <cb@df7cb.de>
Date: 2018-10-12T17:56:55Z
Lists: pgsql-hackers
On 2018-10-12 13:51:53 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > But can't we just do something like:
> 
> > #if defined(HAVE__BUILTIN_CONSTANT_P) && defined(HAVE__VA_ARGS)
> > #define AllocSetContextCreate(parent, name, ...) \
> > 	(StaticAssertExpr(__builtin_constant_p(name), \
> > 					  "memory context names must be constant strings"), \
> > 	 AllocSetContextCreateExtended(parent, name, __VA_ARGS__))
> > #else
> > #define AllocSetContextCreate \
> > 	AllocSetContextCreateExtended
> > #endif
> 
> > The set of compilers that have __builtin_constant_p and not vararg
> > macros got to be about empty.
> 
> Yeah, fair point, and anyway we don't need the StaticAssert to exist
> everywhere.  I'll make it so.

Cool.


> Shall we also backpatch the ALLOCSET_*_SIZES macros as Christoph
> suggested?  I'm not convinced of the usefulness of that, since
> extensions would still have to cope with them not being present
> when building against existing minor releases.

I'd do so. Many extensions are fine just building against a relatively
new minor release. Won't help extension authors in the next few months,
but after that...

Greetings,

Andres Freund


Commits

  1. Back-patch addition of the ALLOCSET_FOO_SIZES macros.

  2. Simplify use of AllocSetContextCreate() wrapper macro.

  3. Rethink MemoryContext creation to improve performance.