"Jason R. Mastaler" <[EMAIL PROTECTED]> writes:

> Good. Some things I'd like to see in CVS before I make the 0.69
> release:
> 
> - Your signal that the virtual domains code is more or less generally
>   usable.

Ugh.  As is, I think it works.  I also think I should remove the code
that passes the -c flag to tmda-inject, since it's not necessary now
that $HOME is set correctly.  tmda-inject will find the user's config
file automatically.  That would simply the VDomainProxy class a bit.

With that change, I'm willing to say the tmda-ofmipd side is done.

I'll probably rip it out later and create a module.  I'm thinking
about a couple of classes, one for standard users and one for virtual
users.  One or the other of those classes would get created very, very
early on in Defaults, probably, and things like Util.gethomedir()
would become config_object.gethomedir(), etc.  One of the
initialization jobs in at least the virtual config class would be to
set the $HOME variable appropriately.

I just need to make sure I don't have any chicken-and-egg problems
with what I know and what I need to know at the time that config class
is created.

The point of this is that suddenly, ~ would mean the right thing for
virtual users in tmda-rfilter, too, which we still really need.

> - Some tmda-cgi bug fixes to the address problems reported on -users
>   recently. Unsure if this will happen as Gre7g seems AWOL at the
>   moment.

Once I have the module I mentioned above in place, I will take a look
at the CGI code to see if I can figure out how to integrate the
virtual domain stuff better with that app.

> - I'm working on a fix to the maildir delivery code to address the
>   weaknesses in the popular TIME.PID.HOST algorithm which has been
>   discussed recently on the various MTA lists. See
>   http://cr.yp.to/proto/maildir.html for djb's updated spec.

I saw that.  Looks like the fallback algorithm, after you don't use
all the parts that you should use but aren't implemented on various
systems, is pretty similar to what maildir writers have been using all
along.  :)

A final note: tmda-ofmipd's -A/--authprog flag is thoroughly broken.
It simply doesn't work.  I think the problem is that David tested with
checkpassword-pam which allows you to specify on what descriptor the
username and password are passed.  This violates the checkpassword
interface as defined by DJB.

Most checkpassword-type programs follow the interface specification to
the letter, which means they expect the username/password on
descriptor 3, not 0.  Most, including DJB's checkpassword, have no way
to specify another descriptor.  Of course, we're passing the
authentication info on descriptor 0, so neither of the checkpassword
implementations in the virtual domain managers work.

I'm working on fixing that today, but I might not get it done.  If you
want to release without this fix, that's fine by me.  I haven't seen
any outcry about it, so I assume no-one is using the -A switch, or if
they tried, they haven't seen fit to report the problem.


Tim
_________________________________________________
tmda-workers mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-workers

Reply via email to