On Tue, Oct 25, 2005 at 10:00:05PM +1000, Philip Yarra wrote:
>
> Yes, but I don't think that I want Slony to use the all-uppercase version,
> since it then fails to match the table as it exists in the database.
Well, yes, but only because you didn't double-quote it when you
created it. This is really a result of the SQL standard deviation by
Postgres.
> I'm looking at it from the principle of least surprise. As a user, it
> surprised me that my table as defined in the SQL script (CREATE TABLE
> ACCOUNTS) was not found by Slony-I when I defined it the same way in
> slon_tools.conf.
That's not quite true. The CREATE TABLE ACCOUNTS is implicitly
folded to lowercase; and you should know that as a PostgreSQL admin.
The slon_tools.conf file, on the other hand, has the name of the
table single-quoted, which ought to be a clue that it will be treated
literally (i.e. case-sensitively).
IIRC, though, the double-quote-everything strategy was added to
handle the case where someone _had_ created a table with mixed case
names. A quick glance at the docs does not produce anything leaping
out at me to indicate for a new admin that these might be case
sensitive. So it does seem the docs could use a patch. Care to
suggest one (and a place for it?)
> So is the quoting to preserve case actually desirable in this function?
For sure: not preserving it would make it impossible to use by people
who _do_ have uppercase table names, to begin with.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The plural of anecdote is not data.
--Roger Brinner
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general
