Re: fast defaults in heap_getattr vs heap_deform_tuple - Mailing list pgsql-hackers

From Christoph Berg
Subject Re: fast defaults in heap_getattr vs heap_deform_tuple
Date
Msg-id 20190206122229.GB10563@msg.df7cb.de
Whole thread Raw
In response to Re: fast defaults in heap_getattr vs heap_deform_tuple  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
> > Sorry I didn't spot this earlier, but... in the backpatch to v11, is the
> > expansion of the trigger tuple really unnecessary now? Since
> > heap_getattr is a macro, C-language triggers that aren't recompiled
> > won't see the default? So perhaps the expansion should be left in?
> 
>        Myon | the list of packages in Debian where heap_getattr appears is postgresql-11, pglogical,
>             | pgextwlist, citus, postgresql-plsh, wal2json, pg-cron, postgresql-rum
> 
> I'm mildly disinclined to re-add the heap_expand_tuple, because it's not
> the only case, and the extensions named above don't seem like they'd
> specifically affected by the trigger path, but I'm not sure.

I'm not following closely enough to say anything about which fix is
the best, but if this isn't a "recompile the world" case, we can get
the above list of packages rebuilt. It would be rather unpleasant to
have to schedule that for 50 packages, though.

Are rebuilt extension binaries compatible with older servers that do
not have the fix yet?

Christoph


pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: Add missing CREATE TABLE IF NOT EXISTS table_name AS EXECUTEquery;
Next
From: Amit Khandekar
Date:
Subject: Re: Pluggable Storage - Andres's take