"Julien Buratto" <kannel 'at' linkas.it> writes: > > Kannel build and control network access with the socket > > interface. It must be possible to "hack" the contents of a TCP > > transmission with enough permissions, but I don't think this is > > what Kannel is doing :). A quick look at the EMI implementation > > seems to confirm that. > > So you think the problem isn't related to the kannel itself ?
Yes. But nothing can beat try and error with hands on the code and the test platform, so you should try to prove me wrong (or right) with more concrete stuff :). > Actually I'm trying to modify the smsc_emi.c file because it > seems to me that it sends sms messages without waiting for the > ack. At the TCP level?? At the upper level, why not.. but at the TCP level hum.. Seems that sending eventually is done in emimsg.c:emimsg_send which performs a conn_write which is doing "normal" socket (actually file descriptor) write AFAIK.. -- Guillaume Cottenceau
