Import sipcc from ikran into media/signaling or media/webrtc/signaling sipcc is licensed under the MPL I believe, so there are no licensing issues. We'll also need to pare it down to the parts needs for the blocked bug (implementing JSEP).

SIPCC Imported - 2012-03-14 19:52 -0700 - into media/webrtc/signaling

capturing patch to allow another ostream converter

md5 patch pushed as https://hg.mozilla.org/projects/alder/rev/b7014697980d

The windows build seems to have two problems. The first is some duplicate defs which are fixed by the patch above. The second is getting chromium_base and mozilla::Logger available to the link of gkmedias.dll.

Note: this works, but there are some evil hacks in there. A polling loop for registration, and a total weirdness about _self=NULL not producing NULL, but _self=0x12345678 does yield 0x12345678

Comment on attachment 614273 [details] [diff] [review] Windows build work: Add lock.h using NSPR Locks, old-style linked_ptrs, etc - works This is the alder trunk patch

The latest patch needs to be back-ported to trunk

Comment on attachment 611926 [details] [diff] [review] Fixes in signaling code from paris_demo f=enda,jesup Review of attachment 611926 [details] [diff] [review]: ----------------------------------------------------------------- A number of issues need to be resolved/explained... Note I didn't go over cpr_dummy_socket.c in detail ::: media/webrtc/signaling/signaling.gyp @@ +474,5 @@ > './src/sipcc/cpr/include/cpr_memory.h', > './src/sipcc/cpr/include/cpr_rand.h', > './src/sipcc/cpr/include/cpr_socket.h', > + './src/sipcc/cpr/dummy/cpr_dummy_socket.h', > + './src/sipcc/cpr/dummy/cpr_dummy_socket.c', Are cpr_dummy_socket.c/etc intended for default/production? (My guess is yes.) ::: media/webrtc/signaling/sipcc-config.mk @@ +31,5 @@ > +# and other provisions required by the GPL or the LGPL. If you do not delete > +# the provisions above, a recipient may use your version of this file under > +# the terms of any one of the MPL, the GPL or the LGPL. > +# > +# ***** END LICENSE BLOCK ***** Not the current "new-code" license block. See http://www.mozilla.org/MPL/headers/ @@ +54,5 @@ > +DEFINES += -DSIP_OS_WINDOWS > +else > +ifeq ($(OS_ARCH),Linux) > +DEFINES += -DSIP_OS_LINUX > +endif Need Android (even if it's a no-op for now), Consider instead using OS_TARGET instead (such as SIP_OS=$OS_TARGET) ::: media/webrtc/signaling/src/sipcc/core/gsm/lsm.c @@ +2148,5 @@ > lsm_debug_entry(call_id, 0, fname); > LSM_DEBUG(DEB_F_PREFIX"called_number= %s\n", DEB_F_PREFIX_ARGS(LSM, fname), called_number); > > +// line = sip_config_get_line_by_called_number(1, called_number); > + line = 1; // EKR This doesn't look right for checkin to default. Add a comment like "WebRTC only supports single registration per instance/peerconnection" (or whatever). Remove the "EKR" ::: media/webrtc/signaling/src/sipcc/core/sipstack/ccsip_messaging.c @@ +3905,5 @@ > > > + config_get_value(CFGID_P2PSIP, &p2psip, sizeof(p2psip)); > + if (p2psip) { > + strncpy(ccb->ReqURI, "sip:dummy@dummy.invalid", sizeof(ccb->ReqURI)); I'm not sure this is the value we'd want to use... though it may not matter. ::: media/webrtc/signaling/src/sipcc/cpr/dummy/cpr_dummy_socket.c @@ +35,5 @@ > + * and other provisions required by the GPL or the LGPL. If you do not delete > + * the provisions above, a recipient may use your version of this file under > + * the terms of any one of the MPL, the GPL or the LGPL. > + * > + * ***** END LICENSE BLOCK ***** */ Maybe wrong license block (if this isn't a modified copy of another file - if it is (and it probably is), this is ok). @@ +96,5 @@ > + int in_use; > +} dummy_socket; > + > +#define MAX_SOCKETS 64 > +static dummy_socket socket_table[MAX_SOCKETS] = { Static arrays like this always concern me. Either it can overflow (usually in an odd instance, causing crashes or security issues), or there's no way it can ever overflow, and it may well be considerably oversized - and there's nothing here calculating the size or describing why 64 was picked (probably because something thought it was "so big it will never be exceeded", in which case tests for hitting the limit may be missing or if there never tested, and if the unstated assumptions were to be changed it could start breaking, breaking security, etc. That said, since it's likely cloned from another blah_socket file, it may not be a new problem. Either deal with this, or more likely file a bug on resolving this (even if it's only by documenting why it's safe).

Comment on attachment 611926 [details] [diff] [review] Fixes in signaling code from paris_demo f=enda,jesup We are not bringing these paris_demo changes into alder/default. The JSEP implementation will not need dummy_socket.