Carl Bowden wrote:
> I had applyed Alans patch of select_db.h for Bug no. 183
> <snip>
> ap_log_rerror(APLOG_MARK, APLOG_WARNING | APLOG_NOERRNO,
> r, "Midgard: switching to %s failed",
> dbc->name);
> - return 0;
> + /* DONT GIVE UP try disconnect and reconnect!*/
> + mgd_close( dbc->handle->mgd);
> + return assert_conn
> + ? mgd_assert_db_connection(rcfg->mgd,
> + dbc->name, dbc->handle->username,
> dbc->handle->password)
> + : 1;
> +
> +
> +
> }
> }
> dbc->handle->current = dbc;
>
> could this create the out of sync thing as its asked for a
> mgd_close
>
> (the bug tracker has reported this as commited to cvs --
> could it be assuming any other changes from cvs?)
No, don't think so. It could be that this is the cause for the 'out of
sync' (although I can't find a direct cause for it), but for this code
to be triggered the connection must have been flawed in the first place.
So still some oddities here. Besides, if I read the MySQL docs right,
we're doing a mysql_store_result right after the mysql_query, so this
Should Never Happen (tm) (I know, I know). May be a silly question, but
you can connect by hand to the other databases OK?
> Also some of the dbs are older 1.4b5(ish?) which dont
> require updating as we are moving
> most of the content to the newer dbs -- but may troubling
> the switching?
I can't think of a circumstance that would trigger this.
Emile
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
