Re: doc review for v13 - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: doc review for v13
Date
Msg-id 20200612074832.GH3362@paquier.xyz
Whole thread Raw
In response to Re: doc review for v13  (Justin Pryzby <pryzby@telsasoft.com>)
List pgsql-hackers
On Thu, Jun 11, 2020 at 09:37:09PM -0500, Justin Pryzby wrote:
> Some new bits,
> And some old ones.

I was looking at this patch set, and 0005 has attracted my attention
here:

> --- a/src/backend/utils/cache/relcache.c
> +++ b/src/backend/utils/cache/relcache.c
> @@ -4240,7 +4240,6 @@ AttrDefaultFetch(Relation relation)
>      HeapTuple    htup;
>      Datum        val;
>      bool        isnull;
> -    int            found;
>      int            i;

Since 16828d5, this variable is indeed unused.  Now, the same commit
has removed the following code:
-   if (found != ndef)
-       elog(WARNING, "%d attrdef record(s) missing for rel %s",
-            ndef - found, RelationGetRelationName(relation));

Should we actually keep this variable and have this sanity check in
place?  It seems to me that it would be good to have that, so as we
can make sure that the number of default attributes cached matches
with the number of defaults actually found when scanning each
attribute.  Adding in CC Andrew as the author of 16828d5 for more
input.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: remove some STATUS_* symbols
Next
From: "李杰(慎追)"
Date:
Subject: 回复:how to create index concurrently on paritioned table