server icon indicating copy to clipboard operation
server copied to clipboard

AMCP - PLAY ON EMPTY LAYER

Open PeterAkakpo opened this issue 6 years ago • 6 comments

Expected Behaviour

anytime a PLAY {channel}-{layer} command is sent to an empty layer, the response should be a PLAY FAILED instead of OK. the above command cannot be considered as being successfully executed command

Current Behaviour

the system returns a status of OK (which shows the validity of the command), but which gives a wrong impression. when you issue PLAY 1-0 filename and filename does not exists, you get a different response (PLAY FAILED) and not OK even though the syntax of the command is valid.

Steps to Reproduce

execute PLAY 1-0 with layer 0 empty (no file loaded)

Environment

  • CasparCG Server version: 2.2.0.4
  • Operating system: Windows 10

Attachments

PeterAkakpo avatar Apr 09 '18 09:04 PeterAkakpo

@ronag Had a comment about this earlier in another issue, which was closed. Robert claims that a OK response only acknowledges that the command is received and valid, not taken effect. This definition is a breaking change in the AMCP protocol, but that is the current implementation, if I understand 2.2 correct. Can you please add some light to this @ronag?

And @Julusian will your work on async handling of scheduled/batched commands help on the blocking nature of commands @ronag is working to resolve?

Edit: Read the post once more and see that I misunderstood the nature of the response. So my comments are not related to this issue, but is still a topic for discussion (perhaps in another thread).

jesperstarkar avatar Apr 09 '18 10:04 jesperstarkar

PLAY 1-0 is an invalid command and should not return OK.

ronag avatar Apr 09 '18 10:04 ronag

No, PLAY is perfectly valid! If you have loaded or paused content on a layer, you can invoke playback with PLAY 1-0

jesperstarkar avatar Apr 09 '18 10:04 jesperstarkar

@jesperstarkar: Yes, you are right. It's LOAD 1-0 which is invalid.

ronag avatar Apr 09 '18 10:04 ronag

I'm not sure what is "correct" for this issue. AMCP is a bit semantically broken as it is right now.

ronag avatar Apr 09 '18 10:04 ronag

The original description of correct behaviour is desired. Everything I mentioned is unrelated, or partially related.

jesperstarkar avatar Apr 09 '18 10:04 jesperstarkar