Re: PostgreSQL
Hamish Moffatt wrote:
>On Fri, Oct 10, 1997 at 11:46:28AM +0100, Oliver Elphick wrote:
>...
>> Does your compilation pass all the regression tests?
>
>I compiled it here fine the other day, but I didn't check
>the output of the tests.
Please do; I expect there to be problems, but it would be nice if there
were none! libc6 is the difficulty, which we have to compile it for
in order to fit in with the policy for hamm.
>
>> There will be a problem with handling the update of users from postgres95
>> to postgresql, because the database must be dumped and reloaded. There wa
>s
>> discussion about this a couple of months back. This was the last I heard
>from
>> Siggy Bentrup who is the current maintainer (not Emanuele as listed in
>> the postgres95 package).
>
I have emailed Siggy Bentrup to ask him whether he is still working on
postgresql.
>Perhaps the package needs to be called postgres62 then;
Call it postgresql; that differentiates the name sufficiently.
> done properly
>users could have both installed and migrate themselves with dump etc.
If we take that route, we should make postgres95 use a different port
and shared memory segment and a different value for PGSQLDATA. This can
probably be done in the preinst script of postgresql. I think it would
be a bad idea to change the default settings for the new package
in order to accommodate the old one; most people will only want to
run the new one.
There are certainly lots of problems in data conversion; we have to do
a pg_dumpall to save the existing database for reloading into postgresql,
but the 1.09 pg_dumpall will lose data about ownerships and privileges.
There is also the problem that table and column names have become
case-insensitive since 1.09; also, the number of SQL reserved words that
postgresql disallows has increased. Therefore we cannot be sure that we can
successfully reload a database. It is also conceivable that someone
is already running two or more separate databases and postmasters on
one machine by manipulating port numbers and so on; an automatic
conversion process is must unlikely to succeed with them.
We have to be extremely careful not to destroy people's data inadvertently.
In the end, it may be best to declare a conflict between postgres95
and postgresql, and leave it to the user to sort it out.
--
Oliver Elphick Oliver.Elphick@lfix.co.uk
Isle of Wight http://lfix.co.uk/oliver
PGP key from public servers; key ID 32B8FAA1
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: