Mailing List Archive

On Fri, 2004-04-16 at 00:58, Anton Nekhoroshih wrote:
> On mine are incorrectly created aliases at creation of users.
> It should seems to me it to be so: (see attach)

I don't really understand how things could go wrong when using the
current version of db_addalias. The user_idnr that goes in is always
checked the functions calling it in user.c, isn't it?

Ilja

On Fri, 2004-04-16 at 00:58, Anton Nekhoroshih wrote:
> On mine are incorrectly created aliases at creation of users.
> It should seems to me it to be so: (see attach)

Ah, I see what you mean now. As a matter of fact, the behaviour as it
was was the right behaviour.

In the deliver_to field of the aliases table, a number is inserted
whenever it's a 'local' alias, pointed directly to a mailbox. Otherwise,
a string is used to denote a different alias (possibly to an external
address).

This is not the most elegant setup, but it works perfectly.

Ilja

÷ ðÔÎ, 16.04.2004, × 16:27, Ilja Booij ÐÉÛÅÔ:

I disagree here below an example not the worker:

./dbmail-adduser a oracle@bkh.inforus.com secret 0 25000000
anton@bkh.inforus.com root@bkh.inforus.com

mail -s test anton@bkh.inforus.com ......

and in log

Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_query: executing
query [SELECT user_idnr FROM users WHERE userid='anton@bkh.inforus.com']
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_free_result
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checking user [anton@bkh.inforus.com] in alias table
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checks [-1]
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_query: executing
query [SELECT deliver_to FROM aliases WHERE lower(alias) =
lower('anton@bkh.inforus.com')]
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
into checking loop
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checking user anton@bkh.inforus.com to 36
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checking user [36]
in alias table
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checks [1]
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_query: executing
query [SELECT deliver_to FROM aliases WHERE lower(alias) = lower('36')]
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
adding [36] to deliver_to address
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_free_result
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_get_result: (row =
1,field = 0) bigger then size of result set
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checking user anton@bkh.inforus.com to (null)
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checking user [(null)] in alias table
Apr 16 01:50:57 bkh dbmail/smtp[12147]: authsql.c,auth_check_user_ext:
checks [1]
Apr 16 01:50:57 bkh dbmail/smtp[12147]: dboracle.c,db_query: executing
query [SELECT deliver_to FROM aliases WHERE lower(alias) =
lower('(null)')]

> On Fri, 2004-04-16 at 00:58, Anton Nekhoroshih wrote:
> > On mine are incorrectly created aliases at creation of users.
> > It should seems to me it to be so: (see attach)
>
> Ah, I see what you mean now. As a matter of fact, the behaviour as it
> was was the right behaviour.
>
> In the deliver_to field of the aliases table, a number is inserted
> whenever it's a 'local' alias, pointed directly to a mailbox. Otherwise,
> a string is used to denote a different alias (possibly to an external
> address).
>
> This is not the most elegant setup, but it works perfectly.
>
> Ilja
>
>
> _______________________________________________
> Dbmail-dev mailing list
> Dbmail-dev@dbmail.org
> http://twister.fastxs.net/mailman/listinfo/dbmail-dev
>

On Fri, 2004-04-16 at 18:43, Anton Nekhoroshih wrote:
> В Птн, 16.04.2004, в 16:27, Ilja Booij пишет:
>
> I disagree here below an example not the worker:
>
> ./dbmail-adduser a oracle@bkh.inforus.com secret 0 25000000
> anton@bkh.inforus.com root@bkh.inforus.com
>
> mail -s test anton@bkh.inforus.com ......

Strange.. can you send the contents (or part of the contents) of your
aliases table?

Ilja

÷ ðÎÄ, 19.04.2004, × 13:42, Ilja Booij ÐÉÛÅÔ:

Before I have altered to myself function db_addalias in field DELIVER_TO
were USER_IDNR and delivery on it alisam did not reach. As soon as all
has altered, as I spoke, all has earned.

> On Fri, 2004-04-16 at 18:43, Anton Nekhoroshih wrote:
> > В Птн, 16.04.2004, в 16:27, Ilja Booij пишет:
> >
> > I disagree here below an example not the worker:
> >
> > ./dbmail-adduser a oracle@bkh.inforus.com secret 0 25000000
> > anton@bkh.inforus.com root@bkh.inforus.com
> >
> > mail -s test anton@bkh.inforus.com ......
>
> Strange.. can you send the contents (or part of the contents) of your
> aliases table?
>
> Ilja
>
>
> _______________________________________________
> Dbmail-dev mailing list
> Dbmail-dev@dbmail.org
> http://twister.fastxs.net/mailman/listinfo/dbmail-dev
>

On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> В Пнд, 19.04.2004, в 13:42, Ilja Booij пишет:
>
> Before I have altered to myself function db_addalias in field DELIVER_TO
> were USER_IDNR and delivery on it alisam did not reach. As soon as all
> has altered, as I spoke, all has earned.

Strange, because the deliver_to field *should* hold the user_idnr. It is
supposed to work that way. It's probably something that's incompatible
with Oracle.

I definitely need an Oracle setup here..

Ilja

You can download a copy from oracle,... It's free for personal / development use.

... John

-----Original Message-----
From: dbmail-dev-bounces@dbmail.org [mailto:dbmail-dev-bounces@dbmail.org] On Behalf Of Ilja Booij
Sent: Wednesday, April 21, 2004 6:53 PM
To: DBMAIL Developers Mailinglist
Subject: Re: [Dbmail-dev] adduser

On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> В Пнд, 19.04.2004, в 13:42, Ilja Booij пишет:
>
> Before I have altered to myself function db_addalias in field
> DELIVER_TO were USER_IDNR and delivery on it alisam did not reach. As
> soon as all has altered, as I spoke, all has earned.

Strange, because the deliver_to field *should* hold the user_idnr. It is supposed to work that way. It's probably something that's incompatible with Oracle.

I definitely need an Oracle setup here..

Ilja


_______________________________________________
Dbmail-dev mailing list
Dbmail-dev@dbmail.org
http://twister.fastxs.net/mailman/listinfo/dbmail-dev

On Wed, 2004-04-21 at 10:57, John Hansen wrote:
> You can download a copy from oracle,... It's free for personal / development use.

I know. I just had some trouble installing it. It wanted to install only
on RHES or UnitedLinux.. I Probably overlooked something.

Ilja

÷ óÒÄ, 21.04.2004, × 12:52, Ilja Booij ÐÉÛÅÔ:
> On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> > В Пнд, 19.04.2004, в 13:42, Ilja Booij пишет:
> >
> > Before I have altered to myself function db_addalias in field DELIVER_TO
> > were USER_IDNR and delivery on it alisam did not reach. As soon as all
> > has altered, as I spoke, all has earned.
>
> Strange, because the deliver_to field *should* hold the user_idnr. It is
> supposed to work that way. It's probably something that's incompatible
> with Oracle.

Then what for the field deliver_to has a format varchar it incorrectly
to use two algorithms on one field let it will be better varchar and
will store real addresses.

>
> I definitely need an Oracle setup here..
>
> Ilja
>
>
> _______________________________________________
> Dbmail-dev mailing list
> Dbmail-dev@dbmail.org
> http://twister.fastxs.net/mailman/listinfo/dbmail-dev
>

On Wednesday 21 April 2004 11:02, Ilja Booij wrote:
> On Wed, 2004-04-21 at 10:57, John Hansen wrote:
> > You can download a copy from oracle,... It's free for personal /
> > development use.
> I know. I just had some trouble installing it. It wanted to install only
> on RHES or UnitedLinux.. I Probably overlooked something.

If I'm not mistaken, there's a debian-install-howto for Oracle somewhere on
the net... Google is your friend :)

Ricardo.

--

--------------------------------------------
Ricardo Kustner
IC&S Linux Professionals
T: +31-(0)30-6355730 F: +31-30-6355731
PGP-key: http://www.ic-s.nl/keys/ricardo.txt

On Wed, 2004-04-21 at 11:03, Anton Nekhoroshih wrote:
> В Срд, 21.04.2004, в 12:52, Ilja Booij пишет:
> > On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> > > п▓ п÷п╫п╢, 19.04.2004, п╡ 13:42, Ilja Booij п©п╦я┬п╣я┌:
> > >
> > > Before I have altered to myself function db_addalias in field DELIVER_TO
> > > were USER_IDNR and delivery on it alisam did not reach. As soon as all
> > > has altered, as I spoke, all has earned.
> >
> > Strange, because the deliver_to field *should* hold the user_idnr. It is
> > supposed to work that way. It's probably something that's incompatible
> > with Oracle.
>
> Then what for the field deliver_to has a format varchar it incorrectly
> to use two algorithms on one field let it will be better varchar and
> will store real addresses.

I agree that the current solution is not elegant. We could probably
change it, but it should not be changed before 2.0

Another strange thing is that your solution is not working here. When
there's a username instead of a user_idnr in deliver_to, dbmail attempts
to do an external forward.

Ilja

÷ óÒÄ, 21.04.2004, × 13:27, Ilja Booij ÐÉÛÅÔ:

I apologize, has found in what a problem, during time recrusive calls it
is lost database statment and as the result db_get_result can return
NULL, and is farther signal 11.

it is necessary to think out something, I initially did not like
functions db_get_result_set and db_set_result_set

:-(

Once again I ask an apology for superfluous noise :-)

> On Wed, 2004-04-21 at 11:03, Anton Nekhoroshih wrote:
> > В Срд, 21.04.2004, в 12:52, Ilja Booij пишет:
> > > On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> > > > п▓ п÷п╫п╢, 19.04.2004, п╡ 13:42, Ilja Booij п©п╦я┬п╣я┌:
> > > >
> > > > Before I have altered to myself function db_addalias in field DELIVER_TO
> > > > were USER_IDNR and delivery on it alisam did not reach. As soon as all
> > > > has altered, as I spoke, all has earned.
> > >
> > > Strange, because the deliver_to field *should* hold the user_idnr. It is
> > > supposed to work that way. It's probably something that's incompatible
> > > with Oracle.
> >
> > Then what for the field deliver_to has a format varchar it incorrectly
> > to use two algorithms on one field let it will be better varchar and
> > will store real addresses.
>
> I agree that the current solution is not elegant. We could probably
> change it, but it should not be changed before 2.0
>
> Another strange thing is that your solution is not working here. When
> there's a username instead of a user_idnr in deliver_to, dbmail attempts
> to do an external forward.
>
> Ilja
>
>
> _______________________________________________
> Dbmail-dev mailing list
> Dbmail-dev@dbmail.org
> http://twister.fastxs.net/mailman/listinfo/dbmail-dev
>

÷ óÒÄ, 21.04.2004, × 13:27, Ilja Booij ÐÉÛÅÔ:

I apologize, has found in what a problem, during time recrusive calls it
is lost database statment and as the result db_get_result can return
NULL, and is farther signal 11.

it is necessary to think out something, I initially did not like
functions db_get_result_set and db_set_result_set

:-(

Once again I ask an apology for superfluous noise :-)

> On Wed, 2004-04-21 at 11:03, Anton Nekhoroshih wrote:
> > В Срд, 21.04.2004, в 12:52, Ilja Booij пишет:
> > > On Mon, 2004-04-19 at 18:25, Anton Nekhoroshih wrote:
> > > > п▓ п÷п╫п╢, 19.04.2004, п╡ 13:42, Ilja Booij п©п╦я┬п╣я┌:
> > > >
> > > > Before I have altered to myself function db_addalias in field DELIVER_TO
> > > > were USER_IDNR and delivery on it alisam did not reach. As soon as all
> > > > has altered, as I spoke, all has earned.
> > >
> > > Strange, because the deliver_to field *should* hold the user_idnr. It is
> > > supposed to work that way. It's probably something that's incompatible
> > > with Oracle.
> >
> > Then what for the field deliver_to has a format varchar it incorrectly
> > to use two algorithms on one field let it will be better varchar and
> > will store real addresses.
>
> I agree that the current solution is not elegant. We could probably
> change it, but it should not be changed before 2.0
>
> Another strange thing is that your solution is not working here. When
> there's a username instead of a user_idnr in deliver_to, dbmail attempts
> to do an external forward.
>
> Ilja
>
>
> _______________________________________________
> Dbmail-dev mailing list
> Dbmail-dev@dbmail.org
> http://twister.fastxs.net/mailman/listinfo/dbmail-dev
>