Oh, wait, it's no good unless db_connect() also gets more connection
handling smart, because essentially it would be writing this:
db_connect();
db_connect();
Which would leak the first a connection into thin air!
Aaron
On Sun, 2 Feb 2003, Aaron Stone wrote:
> Yes, this would work well, but I believe there were a few posts which Roel
> responded to a while back regarding a split of auth and storage databases.
> For example, if someone already had a big mysql user base, but wanted to
> store their messages in posgres?
>
> Aaron
>
>
> On Sun, 2 Feb 2003, Eelco van Beek - IC&S wrote:
>
> > hmm, i think auth* calls that are used by db* calls should be
> > implemented in db.h.
> >
> > i.e:
> >
> > int auth_connect ()
> > {
> > return (db_connect());
> > }
> >
> > Don't shoot me on the prototype, it's just an example.
> >
> > Eelco
> >
> >
> > On zondag, feb 2, 2003, at 03:01 Europe/Amsterdam, Aaron Stone wrote:
> >
> > > In the coming release, I'd very much appreciate if more calls to
> > > auth_connect() were added to parallel calls to db_connect(). Here's a
> > > quick grep of today's CVS to show which files need the added calls:
> > >
> > > [EMAIL PROTECTED] dbmail]$ grep auth_connect *
> > > auth.h:int auth_connect();
> > > injector.c: if (db_connect() != 0 || auth_connect() != 0)
> > > smtp-convert.c: if (db_connect() != 0 || auth_connect() != 0)
> > > user.c: if (auth_connect()==-1)
> > > vut2dbmail.c: if (auth_connect() == -1)
> > >
> > > [EMAIL PROTECTED] dbmail]$ grep db_connect *
> > > db.h:int db_connect();
> > > dbtest.c: db_connect();
> > > injector.c: if (db_connect() != 0 || auth_connect() != 0)
> > > main.c: if (db_connect() < 0)
> > > Binary file main.o matches
> > > maintenance.c: if (db_connect()==-1)
> > > mini-injector.c: if (db_connect() != 0)
> > > raw-convert.c: if (db_connect() != 0)
> > > serverchild.c: if ( db_connect() != 0)
> > > settings.c: if (db_connect()==-1)
> > > smtp-convert.c: if (db_connect() != 0 || auth_connect() != 0)
> > > user.c: if (db_connect()==-1)
> > > vut2dbmail.c: if (db_connect() == -1)
> > >
> > >
> > > Also, for the (default) case of auth_connect() === db_connect(), I
> > > posted
> > > an #ifndef patch a few weeks ago to take care of that. Like this:
> > >
> > > int auth_connect()
> > > {
> > > #ifndef DBMAIL_USE_SAME_CONNECTION
> > > mysql_init(&__auth_conn);
> > > if (mysql_real_connect (&__auth_conn, _auth_host, _auth_user,
> > > _auth_pass, _auth_db, 0, NULL, 0) == NULL)
> > > {
> > > trace(TRACE_ERROR,"auth_connect(): mysql_real_connect failed:
> > > %s",
> > > mysql_error(&__auth_conn));
> > > return -1;
> > > }
> > > #endif
> > > return 0;
> > > }
> > >
> > >
> > > int auth_disconnect()
> > > {
> > > #ifndef DBMAIL_USE_SAME_CONNECTION
> > > mysql_close(&__auth_conn);
> > > #endif
> > > return 0;
> > > }
> > >
> > >
> > > Thanks!
> > > Aaron
> > >
> > > _______________________________________________
> > > Dbmail mailing list
> > > Dbmail@dbmail.org
> > > https://mailman.fastxs.nl/mailman/listinfo/dbmail
> > >
> > _________________________
> > E.J.A. van Beek
> > ICT Manager
> > IC&S
> > T: +31 30 2322878
> > F: +31 30 2322305
> >
> > PGP-key:
> > www.ic-s.nl/keys/eelco.txt
> >
> > _______________________________________________
> > Dbmail mailing list
> > Dbmail@dbmail.org
> > https://mailman.fastxs.nl/mailman/listinfo/dbmail
> >
>
> _______________________________________________
> Dbmail mailing list
> Dbmail@dbmail.org
> https://mailman.fastxs.nl/mailman/listinfo/dbmail
>
