That's good.  Haven't played with postgres lately.  MySQL seemed
easy and quick enough.  I've added a patch that checks the return
code from vopen_smtp_relay().  It used to be void.  I changed the
interpretation such that it needs to return non-zero if the given
IP address wasn't already in the table, and zero on error or the
IP was in the table and just the timestamp was updated.  I don't
know the syntax for postgres.  Would you be able to make that
modification?  I also don't have an oracle Pro C compiler.  I
typically use OCI.  So I don't know how to update the voracle.pc file.
If you or anyone knows how to update that, please send an update.

Also, you don't have to grant DROP privs, just give it CREATE and
it has the auto-create feature built in.  You only need to grant
DROP if you disable the MANY_DOMAINS option, which you probably
don't want to do anyway.

Not sure which database notification daemon you're speaking of.

Thanks,

Brian

  > > Anyone?  There must be some developmental work going on with it or it
  > > wouldn't have been in the tarball.  I am willing to help work on it but I
  > > need some help getting over this initial hurdle (cleanly).
  > 
  > Nice, responsive list.  :-/
  > 
  > Attached is a patch to 5.3.16 which will get vpgsql to actually work (seems 
to 
  > work right for auth and relay, have NOT extensively tested it yet) -- I have 
  > not included the configure script changes since they were really nasty and 
  > I'm sure there is a better way.
  > 
  > ATM the database, user and password are all hardcoded, and I did not change 
  > the existing ability to create the database if it doesn't already exist.  
  > That seems like a really nasty thing to allow -- basically you are granting 
  > the vpopmail database user full access to the vpopmail table where it would 
  > seem more prudent to only allow inserts, updates and deletes. (not 
  > create/drop).  Anyway.
  > 
  > Another nasty is that the database notification daemon will spit its info 
out 
  > to the network (i.e. anyone accessing vchkpw) -- I will include a fix for 
  > that soon as well as some general cleanups and any other bus I find.
  > 
  > I hope to do some more extensive testing, including clear passwords and open 
  > relay in the next few days.  Adding domains/users and authenticating 
  > definately is working though, and the tables seem to be updating correctly 
  > for relay and lastauth.
  > 
  > Regards,
  > Andrew


Reply via email to