On Thu, 2004-11-18 at 21:48 +0100, Alexander Malysh wrote: > Hi, Hi Alex,
> attached you can find a third version of this patch. > what patch does and how it works see: > http://www.mail-archive.com/devel%40kannel.org/msg03831.html > > Changes since second version: > - fixed memleak that key passed to conn_register was not freed > - added conn_register_real that takes callback data and data-destroyer > function pointer > - when conn_destroy/conn_unregister called and Connection has data_destroyer > fn-pointer that will be called in order to destroy callback data > - when conn_register on already registered data called that if old-data != > new-data then data-destroyer function called in order to destroy previous > callback data I'm on a somewhat fast EMI link (about 50 messages/sec throughput) and I have been hitting the max. fd limit recently. Just pulled HEAD from CVS and applied the patch. Here is the minimalistic conf: # Kannel config group = core smsbox-port = 2004 admin-port = 2005 admin-password = "test" dlr-storage = internal group = smsc smsc = emi host = 127.0.0.1 port = 5142 smsc-username = 2747 smsc-password = bulk1234 keepalive = 180 window = 7 group = smsbox bearerbox-host = 127.0.0.1 sendsms-port = 23522 group = sendsms-user username = ucpadp password = 123456 dlr-url = http://127.0.0.1/ucpadp/dlr_url.php?id=%d group = sms-service keyword = default send-sender = 1 assume-plain-text = 1 post-url = "http://127.0.0.1/post" max-messages = 0 accept-x-kannel-headers = 1 concatenation = true This patch causes DLR messages not to be sent and memory usage of smsbox keeps growing. Here is a snip of smsbox output (debug level): 2004-11-19 19:54:31 [17429] [3] INFO: smsbox: Got HTTP request </cgi-bin/sendsms> from <127.0.0.1> 2004-11-19 19:54:31 [17429] [3] INFO: sendsms used by <test> 2004-11-19 19:54:31 [17429] [3] INFO: sendsms sender:<test:1234> (127.0.0.1) to:<xyz> msg:<deneme> 2004-11-19 19:54:31 [17429] [3] DEBUG: message length 6, sending 1 messages 2004-11-19 19:54:31 [17429] [3] DEBUG: Status: 202 Answer: <Sent.> 2004-11-19 19:54:31 [17429] [3] DEBUG: HTTP: Destroying HTTPClient area 0x83654b0. 2004-11-19 19:54:31 [17429] [3] DEBUG: HTTP: Destroying HTTPClient for `127.0.0.1'. 2004-11-19 19:54:31 [17429] [2] DEBUG: HTTP: Creating HTTPClient for `127.0.0.1'. 2004-11-19 19:54:31 [17429] [3] INFO: smsbox: Got HTTP request </cgi-bin/sendsms> from <127.0.0.1> 2004-11-19 19:54:31 [17429] [3] INFO: sendsms used by <test> 2004-11-19 19:54:31 [17429] [3] INFO: sendsms sender:<test:1234> (127.0.0.1) to:<xyz> msg:<deneme> 2004-11-19 19:54:31 [17429] [3] DEBUG: message length 6, sending 1 messages 2004-11-19 19:54:31 [17429] [3] DEBUG: Status: 202 Answer: <Sent.> 2004-11-19 19:54:31 [17429] [3] DEBUG: HTTP: Destroying HTTPClient area 0x83654b0. 2004-11-19 19:54:31 [17429] [3] DEBUG: HTTP: Destroying HTTPClient for `127.0.0.1'. 2004-11-19 19:54:31 [17429] [4] INFO: Starting delivery report <test> from <1234> 2004-11-19 19:54:31 [17429] [4] INFO: Starting delivery report <test> from <1234> 2004-11-19 19:54:31 [17429] [4] INFO: Starting delivery report <test> from <1234> What I expect is, when a delivery report is started, it would dump the HTTP output here, so I'd know the DLR was getting forwarded to dlr-url specified, but it simply doesn't work. -HTH -- Enver ALTIN | http://skyblue.gen.tr/ Software developer @ Parkyeri | http://www.parkyeri.com/
signature.asc
Description: This is a digitally signed message part
