Although I'm not a big user of EKF's reply, my understanding is that it has been broken somehow.
@rmackay9 Yes, the Ardupilot's EKF replay function is broken for a long time, I am very appreciate that you can fix it. It is really helpful. THX
On Thu, 25 Oct 2018, Abner Huang wrote:
@rmackay9 Yes, the Ardupilot's EKF replay function is broken for a long
time, I am very appreciate that you can fix it. It is really helpful. THX
What's broken about it?
Currently Working For Me(tm) on master.
@peterbarker
I follow the Replay link and do as below:
Building Replay
./waf configure --board=linux
./waf build --target=tools/Replay
Using Replay
build/linux/tools/Replay -- --help
Run a log through Replay to generate the plot and EKF data files:
build/linux/tools/Replay -- MyLogFile.BIN
Use mavgraph.py to graph the data
mavgraph.py MyLogFile.BIN GPA.SAcc NKF4.SV
However, when use mavgraph.py to graph the data, It will not come out the graphs but show as below:

BTW: I also use the latest master branch.
@huangwen0907 Are you able to share that logfile?
@huangwen0907 FWIW, I think that logfile may be corrupt; it is probably missing the FMT entry for whatever message number 204 is.
@peterbarker This is the log repaly log
@huangwen0907 OK, thanks, I've reproduced and will have a look.
any updates?
@huangwen0907 Actually, I realised this is a bug I've already fixed. I'm preparing a PR.
@peterbarker nice work, thank you very much.
https://github.com/ArduPilot/ardupilot/pull/9660
Now this seems to complete successfully - generating a valid logfile - but I have not actually looked at the content of that logfile (apart from making sure the NKF messages appear to be present.
I'd appreciate you testing that PR and seeing if it makes sense to you.
I have tested the branch #9660 you posted, now my replay function can work as desired. thank you very much.
Fixed. Thanks for testing, @huangwen0907 !
Most helpful comment
@huangwen0907 Actually, I realised this is a bug I've already fixed. I'm preparing a PR.