OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
On the db_common infrastructure -- design rationale?

From: José Luis Tallón (jltallonadv-solutions.net)
Date: Mon Jan 02 2006 - 09:56:20 CST


Hi and happy new year to all.

    I have just undertaken updating my db_common patches (with more
functionality than the one included in the stock Postfix 2.2/2.3), now
that i have a bit of time.

    I have unfortunately lost my personal archive of this list's
messages for the period Jan-Mar 2005 (somebody care to forward me a
compressed mailbox/maildir containing them? ;) ), and looking in the
archives did not help much, either.

    What i'm looking for is the rationale behind leaving just the
%[sudSUD] expansions (plus the %[0-9] "domain parts" matching), and why
the code uses "flags"( DB_COMMON_{KEY,VALUE}_{USER,DOMAIN} ) to
determine which parts are needed. I understand that this is related to
query optimization (supress queries as soon as possible), but not why it
is done in this way.
    Any and all pointers will be most welcome; Victor's comments --being
the last author himself-- might be most instructive.

Thank you in advance.

Best,
    J.L.