On Thu, Aug 30, 2007 at 08:11:25PM -0400, Jan Wieck wrote:
> You forget that Slony is also supposed to replicate across different 
> Postgres versions. So what are we going to do if DDL with multiple 
> nesting layers of different comment styles is supported by one, but not 
> all Postgres versions inside of a Slony cluster? Somehow convert it to 
> something that is understood by all?

No, we're supposed to support the least common denominator.  If there
isn't one, then I agree tha the user loses.  This is more or less the
pg_dump strategy: use the latest, or die.
 
> Sorry, but Postgres supporting C-style comments in the first place is 
> just plain wrong. Wasting time on shit like this while we don't find the 
> time to implement MERGE, what does that tell you about a project?

It tells me that there are too few developers.  It's not wasting
time.  We can also decide that it's a known issue, in which case (1)
we accept patches and (2) we have a BIG FAT WARNING in the docs about
it.  Neither case seems to be true to me.  

I'm sorry, but dismissing what seems to me to be a totally legitimate
user complaint -- "My replication system breaks on legal PostgreSQL
systems" -- on the grounds that the SQL isn't what you consider "good
SQL style" is totally bogus.  This is a replication system _for
Postgres_, not for some magical sytem out there that happens to be
like Postgres, only isn't.  Postgres accepts C-style comments, with a
bunch of warts that Chris points out.  Don't like that?  Tough. 
That's the system we gots.  Either we document this as a limitation
of our system (and, ideally, provide a robust stripper to remove
offending comments -- yes, I'm brushing the rust off my sed even as
we speak), or we fix it, or we admit that we're a replication system
for PostgresPERFECT, rather than PostgreSQL.

A

-- 
Andrew Sullivan  | ajs at crankycanuck.ca
Everything that happens in the world happens at some place.
		--Jane Jacobs