Re: [PATCH] (sqlbox) sqlbox.patch

omg Rene takes all the responsibility and he has answered before me

I'll think how can i make sqlbox a standalone module.

Rene: Are you with me right? ;)

M

On 6/14/06, Mi Reflejo <[email protected]> wrote:
> err. Rene won't take responsibility. so i'll answer.
>  ;)
>
> On 6/13/06, Stipe Tolj <[email protected]> wrote:
> > Mi Reflejo wrote:
> >
> > > New sqlbox.patch for sqlbox module according to the last gateway update.
> >
> > Hi Martin,
> >
> > thanks a lot for the submission. We appritiate your work on Kannel, and I have
> > to admit that this is a good way to go...
> >
> > Now, some critics anyway (even while Rene is more in charge for sqlbox ;)
> >
> > * Please don't include configure itself in the patchset. It's "generated" code
> > and hence, only configure.in belongs to the patchset. (It's like if we would add
> > .o files to the CVS ;)
> >
>
> First to all; The patch was done by Rene. I don't know exactly why he
> put some things, but i can realize why he put some others.
>
> I agree 100% with you, autoconf must do configure file but I think
> that Rene is including it because users just do "make patch && cd
> ../gateway && ./configure && make install".
>
> > * The Makefile.in change is necessary to let Kannel core build sqlbox. This can
> > be moved out of Kannel core in order to allow sqlbox (as own add-on module) to
> > build itself again Kannel core (gwlib, gw) libs... I could forward you my add-on
> >    sceleton configure.in and Makefile.in to use for such purposes. You could go
> > then and addopt it for sqlbox?
>
> yes, but i don't think that sqlbox can be moved out of Kannel since
> some modification to smsbox need to be made.
> ***
>
> > * Why is
> > diff -u -r1.166 configure.in
> > --- configure.in.orig   2006-06-13 15:27:52.000000000 -0600
> > +++ configure.in        2006-06-13 15:27:15.000000000 -0600
> > @@ -936,14 +936,14 @@
> >           dnl mysql-4.x style
> >           MYSQL_LIBS=""
> >           if $MYSQL_CONFIG --libs_r &>/dev/null ; then
> > -            MYSQL_LIBS=`$MYSQL_CONFIG --libs_r`
> > +            MYSQL_LIBS=`$MYSQL_CONFIG --libs_r | sed -e "s/\x27//g"`
> >               AC_MSG_CHECKING([mysql reentrant libs])
> >               AC_MSG_RESULT([$MYSQL_LIBS])
> >               AC_CHECK_LIB(mysqlclient_r, mysql_init, [ LIBS="$LIBS $MYSQL_LIBS" ],
> >                   [ MYSQL_LIBS="" ], [ $MYSQL_LIBS ])
> >           fi
> >           if test -z "$MYSQL_LIBS" ; then
> > -            MYSQL_LIBS=`$MYSQL_CONFIG --libs`
> > +            MYSQL_LIBS=`$MYSQL_CONFIG --libs | sed -e "s/\x27//g"`
> >               AC_MSG_CHECKING([mysql libs])
> >               AC_MSG_RESULT([$MYSQL_LIBS])
> >               AC_CHECK_LIB(mysqlclient, mysql_init, [ LIBS="$LIBS $MYSQL_LIBS" ],
> >
> > necessary? Was this the issue with the ' signs?
> > If yes, this is of general scope and should go to gateway core itself.
>
> I have seen some very old mysql_config returning '-L/path
> -lmysqlclient -letc' (With single quotes).
>  But i think that isn't necessary anymore.
>
>
> > * Why is config directive 'bearerbox-is-sqlbox' necessary for smsbox.c? I don't
> > see a usage of the added int bearerbox_is_sqlbox in the smsbox.c patchset?
> >
> > Does this mean sqlbox acts towards a "patched" smsbox like a bearerbox? Like this:
> >
> >    user <-HTTP-> smsbox <-> sqlbox <-> bearerbox
> >
> > ?
> >
> well.
> Actually yes, sqlbox acts towards smsbox like a bearerbox.
>
> if bearerbox-is-sqlbox flag is true, smsbox-port is fetched from "sqlbox group".
> (sqlbox listen on that port like bearerbox does and connect to
> bearerbox [smsbox-port from "core group"])
>
> > * sqlbox should have it's own module gwlib/sqlbox-cfg.def and call the
> > cfg_add_hooks() routine I proposed to "load" the semantic checks for the new
> > config group.
> >
> Yes, but sqlbox need to be changed. I'll think how to do it.
> >
> > Hope I wasn't to critic ;)... actually I like the sqlbox itself, but would
> > suggest Rene and you to mangle it the way I used for smppbox, where the module
> > should be as much isolated from patching the core as possible. In the end I came
> > up with the single (still to vote) patch for the cfg_add_hook() patch for
> > gwlib/cfg.[ch], and everything else would be in foobar module dir itself....
> > works perfectly and smooth.
> >
> > Stipe
> >
> > -------------------------------------------------------------------
> > Kölner Landstrasse 419
> > 40589 Düsseldorf, NRW, Germany
> >
> > tolj.org system architecture      Kannel Software Foundation (KSF)
> > http://www.tolj.org/              http://www.kannel.org/
> >
> > mailto:st_{at}_tolj.org           mailto:stolj_{at}_kannel.org
> > -------------------------------------------------------------------
> >
>