> Also, do not use the > pg_dumpall script from v6.0 or everything will be owned by the > postgres super user. Well, couldn't you just do the dump in the postinst then? > Looking into the details, I don't see a clear upgrade path to cover > all possible cases.. I can't even assume the postgres95 postmaster is > running and postgres95 didn't provide a startup script in /etc/init.d > (pgsql will). Maybe just prompt, then do a killall. > Taking into account a comment from Oliver Elphick <olly@lfix.co.uk> as > well as a private email from David R Baker <dabaker@InfoAve.Net> on > the same subject I will look into letting pgsql coexist with its > predecessor postgres95. This will necessitate changing the port on > wich the pgsql postmaster is listening as well as using different data > and library directories and diverting postgres95 executables that are > replaced by upgraded pgsql versions. That's a bad idea. Why would anybody want to have postgres95 hanging around? And definitely don't change the port for the postmaster - or any client/server apps and drivers that want to communicate to pgsql will have to be changed too. My guess is that hardly anyone is using the postgres95 package - so don't spend too much effort trying to do everything automatically. > > Another comment, why is it called "pgsql"? Wouldn't it be less cryptic > > to use the long name "postgresql"? > > Simply a matter of taste or call it sheer lazyness. I like "pgsql". It's shorter and has less vowels. :-) Cheers, - Jim
Attachment:
pgpRrqHP0tniW.pgp
Description: PGP signature