+1 from me. can you please commit it?
Thanks,
Alex
Am 31.08.2009 um 12:27 schrieb Alejandro Guerrieri:
Ok, here it is.
<kannel-outgoing-queue-limit.diff>
Regards,
--
Alejandro Guerrieri
[email protected]
On 31/08/2009, at 12:02, Alexander Malysh wrote:
yes, I think bearerbox.h is a good place...
Am 31.08.2009 um 12:01 schrieb Alejandro Guerrieri:
Ok, where do you suggest to place the macro definition? bearerbox.h?
Regards,
--
Alejandro Guerrieri
[email protected]
On 31/08/2009, at 11:50, Alexander Malysh wrote:
Hi,
DEFAULT_OUTGOING_SMS_QLENGTH please :)
Otherwise some people may think about incoming queue too.
Thanks,
Alex
Am 31.08.2009 um 11:30 schrieb Alejandro Guerrieri:
Well, then I'd do it directly when loading the value on
max_outgoing_sms_qlength instead of passing -1 and changing it
to 1M later on.
if (cfg_get_integer(&max_outgoing_sms_qlength, grp,
octstr_imm("sms-outgoing-queue-
limit")) == -1)
max_outgoing_sms_qlength = -1;
==>
#define DEFAULT_SMS_QLENGTH 1000000 (on bearerbox.h?)
....
if (cfg_get_integer(&max_outgoing_sms_qlength, grp,
octstr_imm("sms-outgoing-queue-
limit")) == -1)
max_outgoing_sms_qlength = DEFAULT_SMS_QLENGTH;
I'll prepare a doc patch in the meantime.
Regards,
--
Alejandro Guerrieri
[email protected]
On 31/08/2009, at 11:11, Alexander Malysh wrote:
Hi Alex,
sorry for delay... Yes you are right, -1 meanst not infitine
but large enough IMO. We should fix docs but not the code.
Because if someone want to allow bigger queue then he can set
it in config.
Thanks,
Alex
Am 29.08.2009 um 12:40 schrieb Alejandro Guerrieri:
Alex, can you confirm? If that's the case, IMO we should
either fix the documentation or change the code. On a bulk-
sending scenario 1M is far from being considered infinite.
Regards,
--
Alejandro Guerrieri
[email protected]
On 29/08/2009, at 10:27, Nikos Balkanas wrote:
Hi,
It is common for some configuration variables to be converted
to more manageable values in the code, depending on function
calls. Seems 1000000 is considered large enough in this
implementation to be practically infinite.
@Alex M: Could you please check the Reply-To field in the
lists (devel&users). Lately it is all wrong.
BR,
Nikos
----- Original Message -----
From: Alejandro Guerrieri
To: Kannel Devel
Sent: Saturday, August 29, 2009 2:16 AM
Subject: Queue length
Hi,
I was checking the code that handles the outgoing queue and
I've came into this that got me confused.
On gw/bb_smsconn.c, around line 1170:
/*
* if global queue not empty then 20% reserved for old msgs
* and 80% for new msgs. So we can guarantee that old
msgs find
* place in the SMSC's queue.
*/
if (max_outgoing_sms_qlength > 0 &&
gwlist_len(outgoing_sms) > 0) {
max_queue = (resend ? max_outgoing_sms_qlength :
max_outgoing_sms_qlength * 0.8);
}
else
max_queue = (max_outgoing_sms_qlength > 0 ?
max_outgoing_sms_qlength : 1000000);
However, the documentation states:
sms-outgoing-queue-limit number of messages Set maximum size
of outgoing message queue. After number of messages has hit
this value, Kannel began to discard them. Value -1 default,
means that the queue of infinite length is accepted.
Variable max_outgoing_sms_qlength gets defined on gw/
bearerbox.c line 494:
if (cfg_get_integer(&max_outgoing_sms_qlength, grp,
octstr_imm("sms-outgoing-
queue-limit")) == -1)
max_outgoing_sms_qlength = -1;
So, assuming I don't define it (gets loaded with -1), the
"else" condition would be true and since
max_outgoing_sms_qlength < 0, max_queue would be 1000000 right?
In other words: is the documentation wrong and -1 means
1000000 instead of infinite?
Regards,
--
Alejandro Guerrieri
[email protected]
