(Bug) Deser function in MavFrame/MavHeader not working properly

[IhsenBouallegue]

While using rust mavlink to serialize MavFrames and send them over a LoRa link, I noticed some weird inconsistencies. I noticed a weird problem. Not sure if I am the only one having it. Here are my findings, if you can point me I will be happy to fix it and open a PR.

All the values inside the MavMessage are wrong. But not random! so it's not a problem with the LoRa link. I also printed the buffers out and confirmed that they are identical. The problem happens when deserialising. I noticed something peculiar 20 is 0000000000010100 while 5120 is 0001010000000000 1792 is 0000011100000000 while 0000000000000111 It is consistent. At first, I suspected it has something to do with big/little endian. But maybe the bits don't quite align for that? Not sure what it can be, because I can't find the code responsible for serializing and deserializing the message anywhere in the repo. Is it being imported from somewhere? I would appreciate any hint 😃

[patrickelectric]

Merged and available on latest release: https://github.com/mavlink/rust-mavlink/releases/tag/0.12.1

[IhsenBouallegue]

hey @patrickelectric thank you for releasing this fast. But this issue, while related to PR https://github.com/mavlink/rust-mavlink/pull/204, is not the same. That PR fixes another related problem (I didn't bother creating an issue for it, because it was straightforward). As far as I understood, the problem with the parsing of the MavMessage (discussed here) was not fixed yet. So this issue remains open. Let me know if I missed something.

[patrickelectric]

Hi @IhsenBouallegue, sorry for closing the issue early.

It's a weird problem, to be sure, what kind of hardware is sending this messages ? is it x86 ? is it ARM ? It's probably not an issue with the LoRa link, but it's the first time that I'm seeing such problem.

[patrickelectric]

And how the companion part is running ? Is it using rust-mavlink as well ?

[IhsenBouallegue]

@patrickelectric I threw together a reproduction repo (no lora, just ser and then deser, and still getting the same output) https://github.com/IhsenBouallegue/mavlink-rust-test This also means it has nothing to do with the hardware. But here is what I tried it on:

• on companion computer and ground station (Raspberry pi 4) running Pi OS 64Bit and compiling with target arch aarch64-unknown-linux-gnu
• Windows 11 compiling with x86_64-pc-windows-msvc. Here is the output (basically the same)

  MavFrame {
  header: MavHeader {
      system_id: 1,
      component_id: 1,
      sequence: 84,
  },
  msg: LINK_NODE_STATUS(
      LINK_NODE_STATUS_DATA {
          timestamp: 92197916,
          tx_rate: 53,
          rx_rate: 20,
          messages_sent: 47,
          messages_received: 21,
          messages_lost: 0,
          rx_parse_err: 0,
          tx_overflows: 0,
          rx_overflows: 0,
          tx_buf: 0,
          rx_buf: 0,
      },
  ),
  protocol_version: V2,
  }
  MavFrame {
  header: MavHeader {
      system_id: 1,
      component_id: 1,
      sequence: 84,
  },
  msg: LINK_NODE_STATUS(
      LINK_NODE_STATUS_DATA {
          timestamp: 23602666496,
          tx_rate: 13568,
          rx_rate: 5120,
          messages_sent: 12032,
          messages_received: 5376,
          messages_lost: 0,
          rx_parse_err: 0,
          tx_overflows: 0,
          rx_overflows: 0,
          tx_buf: 0,
          rx_buf: 0,
      },
  ),
  protocol_version: V2,
  }

  (On a good note, the issue with the sequence is fixed 😄)

[patrickelectric]

Thanks for creating your repository @IhsenBouallegue.

It should be fixed by #206 now, I just used in your repository and it indeed solves the problem.

I'll add your example as a test.

[IhsenBouallegue]

Hey @patrickelectric, most welcome! Glad to see it fixed. Would you be so kind and do a quick hotfix release, so I can use it in my full repo?

[patrickelectric]

Hey @patrickelectric, most welcome! Glad to see it fixed. Would you be so kind and do a quick hotfix release, so I can use it in my full repo?

Soon after merging a second test, I'll update you.

[patrickelectric]

@IhsenBouallegue tag done, should be available soon.

[IhsenBouallegue]

@patrickelectric Yup it works! Can't thank you enough :D You literally saved my thesis 😆 You, my friend, will be going into the acknowledgements 🏆

[patrickelectric]

@patrickelectric Yup it works! Can't thank you enough :D You literally saved my thesis 😆 You, my friend, will be going into the acknowledgements 🏆

wow, Thank you :laughing: