Hi,

> Here is the V96 version which:

Thanks for updating the patch. I reviewed v96-0001 and LGTM. I have one comment
for codes in HEAD, let me share just in case.

```
        /* Create conflict log table. */
        relid = heap_create_with_catalog(relname,
                                                                         
PG_CONFLICT_NAMESPACE,
                                                                         0, /* 
tablespace */ 
...
```

In create_conflict_log_table() the tablespace is specified as 0, it bypasses a 
GUC
default_tablespace. Is it intentional? The parameter can be resolved by
GetDefaultTablespace().

Even if we retain the behavior, it can be InvalidOid

Best regards,
Hayato Kuroda
FUJITSU LIMITED

Reply via email to