telemetryDeviceDumper: problems in some .mat files saved

[Nicogene]

In #87 we tried to dump some quantities for the first time using telemetryDeviceDumper and we obtained this list of files:

test_telemetry_old.zip

We found some issues and they have been fixed in #105 #104 #103 . When I analyzed the data I probably forgot to try to load all the data.
If I try to load all the data I get a problem for the first one

Error using load
Cannot read file C:\Users\ngenesio\Downloads\test_telemetry\test_telemetry_1615543915.280656.mat.

@AlexAntn today tried again to run the same experiment(demoYoga) and dumping with the latest tdd and he got this dataset

test_telemetry_new.zip

And for this run I get errors in loading the first two

>> load('test_telemetry_1616666181.080718.mat')
Error using load
Cannot read file C:\Users\ngenesio\Downloads\test_telemetry\test_telemetry_1616666181.080718.mat.
 
>> load('test_telemetry_1616666240.701385.mat')
Error using load
Cannot read file C:\Users\ngenesio\Downloads\test_telemetry\test_telemetry_1616666240.701385.mat.

Probably they are corrupted?
Any Idea @S-Dafarra on how to check what is inside a mat file without loading through matlab/octave?

[Nicogene]

Hold on, the new data has to be retake, we will update this issue soon!

[S-Dafarra]

Just brainstorming:

• I suppose that the third file works correctly, right?
• Can you try to avoid adding a point in the file name? Just to be sure that it is not a weird problem related to the detection of the extension.
• Are you storing strings with UTF8 chars? There is an issue on previous matio versions: UTF-8 string in level 5 mat file fails to load or is corrupted tbeu/matio#164
• Can you add an assert like assert(linearMatrix.size() == (buffInfo.m_dimensions[0] * buffInfo.m_dimensions[1] * num_timesteps) https://github.com/robotology-playground/yarp-telemetry/blob/327cc2901b266198b72d3f5b1ba0a5d3edec0686/src/libYARP_telemetry/src/yarp/telemetry/BufferManager.h#L331 to make sure that we are not reading garbage?

[Nicogene]

Actually, we forgot to install and we used the old version of the device(without the recent changes that fixed the concurrencies)
Now we have a different problem #110

[Nicogene]

test_telemetry_fixed.zip

These are 4 files saved by the tdd fixed (#111). These have been taken attaching to the icub gazebo control boards. The robot was not moving, this is why velocities and acceleration are 0.
Anyhow, I can load all files correctly and now are all ~ of the same size as expected(~ 9 MB for 2 min of acquisition) except for the last one that has been save with what we had, on closure.

With the previous version of tdd the first files were smaller (~ 2MB) and probably corrupted for #110

I think we can say that this have been fixed, closing