Not a bug report

Hi guys, and seasons greetings. I haven't been on the forum much lately as I haven't had any issues or been able to assist in any way.
Until now that is???
I have been using, testing, playing with www.linuxcnc.org/testing-stretch-rtpreempt/ from pretty much day one of it being there and always kept up with the latest version.
I think it currently dishes out version 2.7.14 but I could be wrong, whatever, it works flawlessly for me and I love it.
Recently, in fact, yesterday I did an apt-get upgrade for another reason and it pulled down version 2.8.0-pre1-4081-g48726d6ea.
Fine I thought, I have used the testing 2.8 versions before and known no difference, all worked well.
Unfortunately for me, this version plays havoc with my A axis, "joint3 following error" is the warning but no specific error is given.
I have attached a pic of the line it stalls on and I have to skip a line to resume.
I have also added sample files if anyone wants to replicate the issue. File 1.ngc works perfectly but when wrapped for rotary as in 1_wrap.ngc it gives muchly grief.
And this is an issue through all my wrap files which work perfectly well in all previous versions.
Any thoughts other than going back to the install version which of course I have done?

It might be useful to plot the following error with halscope to get a handle on whats going wrong here.

Like I said the message points to no specific error for tracing.
I am not looking to fault find but having found one, bringing it to attention without knowing where to file a bug report and supplying the means to replicate the error for anyone who does wish to trace it.

Sorry if that comes across as impertinent, it's not intended to be.

A following error _is_ a specific plottable error

That is you can trigger halscope on the following error to get
some idea of whats going wrong.

If you "halscope" joint.3.f-error and trigger on joint.3.f-errored
you should get a lot more information than just the fault popup.
Its probably useful to plot joint.3.vel-cmd at the same time

Thank you for that information. I won't take the trouble to trace errors It's easier to use the stable version which works just fine.
Whatever the issue is, it's intentional as today's upgrade also has it.

OK I'm a bit confused here, Your hal and ini files are LinuxCNC 2.7 version files (using Axis not joints)

Oops, a senior moment between machines.
Of course, the upgrade overwrites them, here are the new ones.

Your old 2.7 config has a MinFerror of 0.25 for A
The new 2.8 config has a MinFerror of 0.10 for A

If you change the value, you will have the same behavior in both configs!

Norbert

Thanks but actually it made no difference whatsoever that was only where I had previously been changing due to reading false information about the error.
The behavior remains the same regardless of those values but from the little research I had done the advice was to change them.

upgreating to 2.8 is not recomended is this a real Gantry so a need for double axis homing
why dident you stay with 2.7.14