You could "use the source" Lourens ?
In gwlib/cfg.[ch] looks like config file is loaded into
generic structures so
e.g. adding to dlr-db group or mysql-db group:
delete-dlrs = 0
or:
delete-dlrs = 1
is already supported?
I could be wrong as I just glanced at it but I don't think so.
I've a preference for 0/1 as opposed to false/true
Then you've only got to find out where dlrs are deleted ...
gw/dlr.[ch] might be a good place to start? :)
look here in dlr.c in dlr_init:
/* check which DLR storage type we are using */
grp = cfg_get_single_group(cfg, octstr_imm("core"));
dlr_type = cfg_get(grp, octstr_imm("dlr-storage"));
looks like it is instantiating structure for accessing values from
"core" group and getting value of "dlr-storage" parameter.
that dlr_type is a global veriable (declared in dlr.h) the grp
is local to dlr_init
poking around more:
grep "log-level" gw/*.c
/home/jamesc/src/topupclub/gateway-1.2.1/gw/bearerbox.c: if
(cfg_get_integer(&loglevel, grp, octstr_imm("log-level")) == -1)
/home/jamesc/src/topupclub/gateway-1.2.1/gw/smsbox.c: cfg_get_integer(&lvl, grp,
octstr_imm("log-level"));
this is how we can get integer values as opposed to string values
look again in dlr.c in dlr_find_sdb and dlr_find_mysql
one of the last things each does is this (more or less equivalent):
if ((typ & DLR_BUFFERED) && ((dlr_mask & DLR_SUCCESS) || (dlr_mask & DLR_FAIL))) {
debug("dlr.sdb", 0, "DLR not deleted because we wait on more reports");
} else {
debug("dlr.sdb", 0, "removing DLR from database");
sql = octstr_format("DELETE FROM %s WHERE %s='%s' AND %s='%s' LIMIT 1",
octstr_get_cstr(table), octstr_get_cstr(field_smsc),
smsc, octstr_get_cstr(field_ts), ts);
* connect to db and execute delete *
}
if we change this in both places to :
if ((typ & DLR_BUFFERED) && ((dlr_mask & DLR_SUCCESS) || (dlr_mask & DLR_FAIL))) {
debug("dlr.sdb", 0, "DLR not deleted because we wait on more reports");
+ } else if (delete_dlrs == 0) {
+ debug("dlr.sdb", 0, "DLR not deleted because delete_dlrs = 0 in .conf");
} else {
debug("dlr.sdb", 0, "removing DLR from database");
sql = octstr_format("DELETE FROM %s WHERE %s='%s' AND %s='%s' LIMIT 1",
octstr_get_cstr(table), octstr_get_cstr(field_smsc),
smsc, octstr_get_cstr(field_ts), ts);
* connect to db and execute delete *
}
AND add global var delete_dlrs to dlr.h
Mutex *dlr_mutex;
Octstr *dlr_type;
+ static long delete_dlrs;
AND get value of delete_dlrs in dlr_init
/* check which DLR storage type we are using */
grp = cfg_get_single_group(cfg, octstr_imm("core"));
dlr_type = cfg_get(grp, octstr_imm("dlr-storage"));
+ /* delete dlrs 1 true, 0 false */
+ cfg_get_integer(&delete_dlrs, grp, octstr_imm("delete-dlrs"));
+ /* -1 (parameter not specified - default false) */
+ if (delete_dlrs == -1) delete_dlrs = 0;
Then ... those 9 lines could add the feature you need?
They compile okay and boxes run but I'm not going to test further - today anyway,
This is the shoddiest glance at the code and guess as to what
works but I hope it gets the general idea across ?
James.
On Tue, May 18, 2004 at 06:39:32PM +0530, Sahil Aggarwal wrote:
>
> i never tried this
> but wht if u revoke delete rights from the user kannel is using to
> delete rows from the dlr table?
> On Tue, 2004-05-18 at 18:24, Lourens Janse van Rensburg wrote:
>
> I agree that it's a dirty hack, but it's been working
> in a production environment for several months now
> without any problems.
>
> I think the most elegant solution would be to add an
> option to the .conf file to indicate that Kannel
> should in fact NOT delete the DLRs from its DLR table
> when the messages have been delivered (or not
> delivered).
>
> Something like:
> group = dlr-db
> id = ...
> table = ...
> delete-dlrs = false
>
> What is the procedure to bring this request to the
> developers' attention ? I haven't posted to the
> developers list before.
>
> Let's help to make Kannel even better.
>
> Lourens...
>
>
>
> -Regards
>
> Sahil Aggarwal
>
> ValueFirst Messaging Pvt. Ltd.
>
> www.vfirst.com
--
http://web.tiscali.it/cedric/jacques_brel_show.html
http://www.thehelix.ie/2004_06_02_Jacques_Brel.html
June Wed/Thur 2/3, Jacques Brel - Ced & Stella
http://www.greenpeace.org/international_en/features/details?item_id=327510
Greenpeace tree-huggers indicted by U.S..
Brazilian mahogany barons curtailing free speech. Grrr,
The following statement is false.
The preceding statement is true. - Douglas Hofstadter
I, James row worse maji.
http://www.bcoleman.com my brother Bill Coleman
He's a bit good! Launching his debut CD right now.
software patents (coming soon to an EU near you) bad for users, for developers
good for monopolistic innovation stifling big business
http://swpat.ffii.org/
beware entering overwhelming twisty maze of political legalese
