Re: User defined data types in Logical Replication

Hi,

It seems to be in the middle of discussion, but I became a reviewer of
this problem several days ago so I've tested the latest patch
'fix_slot_store_error_callback_v12.patch'.

I reproduced the below ERROR by inserting elog() to INPUT function
of CREATE TYPE, and confirmed that above patch solves the problem.

 >>  ERROR:  XX000: cache lookup failed for type XXXXX

I also ran make check-world and there was no error.

Is the only remaining topic memory leaks?


On 2018/01/09 17:22, Masahiko Sawada wrote:

 >> I suppose it's not a common use pattern, but it'd be good
 >> to avoid everlasting memleaks.
 >
 > I agree. Can we remove entry from hash table in the callbacks instead
 > of setting InvalidOid when invalidate caches?


-- 
Atsushi Torikoshi
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center