DICEA "supports" chan_ss7 in a very bad way. There is a "FIXED" version of it, and free version INTENTIONALLY filled and left with bugs in different places. One of the "intentional" bugs - accidental one way audio after a while if usage on certain CIC's , another - improper attaching to the /dev/zap and improper asiggn of the variable value in case of someone would want to use zaptel new style zap addressing (though very simple for anyone with basic C knowledge to fix, patch was provide in the list a long ago) - so it becomes clear that that pieces of code were left intentionally wrong to keep certain number of users out from the functionality they need, and trying to force them for commercial support (for changing those 2 lines of code they asked me for ~2000Euro for human labour) I did not do detail analysis of the code, but since there are already such an approach - I afraid there is more black holes and undiscovered bugs. Instead of making profit on proper support of the users, they prefer to release broken version and make money on those who tired from bugs, they discover while using, and leave the code intentionally not broken. It's already questionable if DICEA doing this legally correct: this may be only legally correct if they got (purchased or whatever) from SIFIRA, otherwise they even break GPL. Further more, they use parts of ZAPTEL code (echo cancel, jitter and maybe something else) - considering this - it's clear GPL breakage - in the view of releasing "closed" version of chan_ss7. Would be good to see their response on the said above, but in fact they very rarely respond at all, even concerning discovered bugs - since they do not intend to fix them. The best way would be to fully switch to libss7, and improve it to add missing functionality. I personally now testing libss7 and if satisfied with stability and performance will switch all my links over to it form chan_ss7 ------
|