Attention is currently required from: pespin.

dexter has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-mgw/+/34349?usp=email )

Change subject: mgcp_client_fsm: allow the same codec multiple times in ptmap ......................................................................

Patch Set 3:

(4 comments)

Commit Message:

https://gerrit.osmocom.org/c/osmo-mgw/+/34349/comment/0bb31340_ea6347a7 PS2, Line 10: he which to used and optionally a payload type map (ptmap) that is used

"he which" makes no sense :)

Done

Patchset:

PS2:

After looking at the whole patch, I have the feeling this patch is a HACK and not a real fix. […]

The way how ptmap works is not changed by this patch. It still works as expected when used with unique codec/pt pairs. All we do now is to allow the same codec appearing multiple times under the assumption that the order matches to order we have in the codecs[] array. I think this is an elegant fix since it does not change the API interface. It only extends the way how ptmap is used a little bit. Since we do not have to change the struct in any way we do not break the ABI, nor the API.

File include/osmocom/mgcp_client/mgcp_client_fsm.h:

https://gerrit.osmocom.org/c/osmo-mgw/+/34349/comment/f73e3b0f_aedc8b71 PS2, Line 39: * codecs array above). In case the same codec type (enum mgcp_codecs) is appearing multiple times in codecs.

This sentence is wrong and I'm not following it: "In case ... […]

Done

File src/libosmo-mgcp-client/mgcp_client.c:

https://gerrit.osmocom.org/c/osmo-mgw/+/34349/comment/f61bce52_98b8cf60 PS2, Line 152: unsigned int map_codec_to_pt2(bool *ptmap_used, const struct ptmap *ptmap,

I'd say better move the "ptmap_used" arg to the end, since it's the only in-out param (so when readi […]

Are you sure? Our convention is to have the output parameters first and then everything that is input. Or do I miss something here?