Add blackbox device FILE for SITL build

[bartslinger]

To make logging for software-in-the-loop (SITL) easier, I added BLACKBOX_DEVICE_FILE as output option for Blackbox. This creates a file YYYY_MM_DD_HHMMSS.TXT when the controller is armed.

I tested this on 7.1.2, but on master I'm running into other issues which prevent me from arming the drone.

[mmosca]

set file_handler to NULL after closing

[bartslinger]

Good point, thanks

[bartslinger]

Hmm still some problems with decoding the logfiles, please hold :)

[bartslinger]

I'm not sure what's happening here. With 7.1.2, there seem to be +- 3 frames that fail to decode when using blackbox_decoder. It seems to be at the very beginning of the file because it also happens when I log for just one second.
I guess this is fine?

However when I run on master, everything fails to decode. I guess this is probably because the log messages have been updated and are not backwards compatible with the version of blackbox_decode? If that's the case, the problem is in the decoder and this pull request should be fine.

./blackbox_decode ~/git/inav/build_SITL/2024_07_12_233059.TXT
Decoding log '/Users/bart/git/inav/build_SITL/2024_07_12_233059.TXT' to '/Users/bart/git/inav/build_SITL/2024_07_12_233059.01.csv'...

Log 1 of 1Statistics
S frames      22   37.0 bytes avg      814 bytes total
Frames        0
Data rate: Unknown, no timing information available.

4129 frames failed to decode, rendering 4080 loop iterations unreadable. 1 iterations are missing in total (0ms, 100.00%)

Generated by INAV blackbox_decode 7.1.0 6eb2fe7

[stronnag]

The current BBL decoder works fine with 8.0.0 logs. The problem is most likely caused by the PR.
There is a inav8_rcmodes that supports the new INAV 8 RC modes, but I don't think that is your problem.

...

Checked: INAV 8.0.0 BBL generated by a FC are correctly read by either blackbox_decode master or inav8_rcmodes branches.

[stronnag]

I suspect that the SITL loop timing completely screws the log timing and framing. This was my conclusion when I attempted the same thing earlier (and paused the attempt as I didn't have time to look further).

[bartslinger]

But that problem is somewhere else in the SITL code then, right? And it only surfaces now because this is the only viable way to get blackbox logs out of SITL.

I'm willing to go a little bit into that rabbit hole, but I could use some pointers on what has changed in SITL since 7.1.2.

[bartslinger]

The problem is caused by the number of servo's being increased to 18 b6f2976

That is incompatible with max 16 servo's in the parser https://github.com/iNavFlight/blackbox-tools/blob/master/src/parser.h#L20

When I set it back to 16, the logfile produced by this pull request seems to be fine!

[stronnag]

Good catch. Keeping the tools up to date with the FC is a nightmare.
Can you please raise either an issue or PR in the blackbox tools repo?

[bartslinger]

Good catch. Keeping the tools up to date with the FC is a nightmare. Can you please raise either an issue or PR in the blackbox tools repo?

iNavFlight/blackbox-tools#87

[mmosca]

Good catch. Keeping the tools up to date with the FC is a nightmare. Can you please raise either an issue or PR in the blackbox tools repo?

How can we handle this better?

[stronnag]

Good catch. Keeping the tools up to date with the FC is a nightmare. Can you please raise either an issue or PR in the blackbox tools repo?

How can we handle this better?

Don't know. I'm pretty well attuned to changes in things like rc_modes and remember to update those.
But it's easy for everyone (including the BBL tools maintainer) to miss other things. I saw your servo PR and (mea culpa) thought nothing more of it.