This doesn't look good. If we throw a WARNING, why do we not insert anything into pg_description. Seems we should throw an error, or do the insert with a warning.
It essentially makes the behavior deprecated and allows dumps to be restored properly (without the extra-database comments.) Here's a thread on the topic:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=bf1obi%24u7k%241%40FreeBSD.csie.NCTU.edu.tw&rnum=7&prev=/groups%3Fq%3D%2522COMMENT%2BON%2BDATABASE%2522%26ie%3DUTF-8%26oe%3DUTF-8%26hl%3Den
I don't know if Rod has plans to change attempts to COMMENT ON non-local databases to an ERROR in 7.5 or not. It was my fault from the beginning - but once I'd implemented COMMENT ON for tables and columns I just couldn't stop... :-)
Mike Mascari
Mike Mascari wrote:..
The comments are stored only in the database's pg_description where the COMMENT ON took place. This caused dump/reload problems. I believe Rod Taylor added the new warning:
[EMAIL PROTECTED] select count(*) from pg_description; count ------- 1541 (1 row)
[EMAIL PROTECTED] COMMENT ON DATABASE test IS 'Hello'; WARNING: database comments may only be applied to the current database COMMENT
[EMAIL PROTECTED] select count(*) from pg_description; count ------- 1541 (1 row)
---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend
