Changing material but spot not working.. maybe a bug ot it's me?

I opened another file: the error is consistent:

First physical part is OK.

From the second physical part (after line 92 in my file) to the last one, ALL center spots tiny moves get changed to a bigger move, causing the error.
I really do not understand how it is possible, it's a really strange behaviour.

And please note in the first part the tiny moves are correct. From the second physical part only they are wrong, changed.

It looks to me like an overflowing variable/memory location, transforming a floating point number in a negative one or something like that.. but why it is not happening to the first Center Spots? It is happening only if before the Center Spots I already cut normal lines.

I installed LinuxCNC by the ISO, ver 2.9.1 some month ago (but I did not cut nested parts with nested center spot). I "apt updated/upgraded" not more 10 days ago. So LinuxCNC got upgraded so I'm now using QtPlasmaC v2.9.236.301

If no one get my error I will reinstall new from the ISO and hope the problem goes away.



Thanks

Thanks again, so I sure need to upgrade again.

I installed using the ISO, but I will check other methods.

Here i posted a video of the error if someone want to check it, (and sorry, my spoken English is a lot worse than my written one !!)



Thanks again for helping me. I hope upgrading the problem will go away,
Alberto.

In the edit all tiny moves are correct:
 

I saved it, rewriting the same file, but in the code display it remains wrong...

Some days ago I made #sudo apt upgrade && sudo apt update, so some libraries were updated, as well as LinuxCNC itself.
Maybe an updated library used in the code display and not in the edit display has something wrong or not compatible...

No, I only have custom.hal and custom_postgui.hal.

The second one is empty (only a comment inside).

Today I will do a fresh install again and check if the problem goes away.

 

I see the clock is 3 hours late (no internet connection) but I don't think could be the cause of the problem...

So, follow up, I discovered something, maybe this could happen in other cases too, so, maybe it's of interest:

resume: if I change the preceding line of the tiny move from G91 to G90 (I said reversed in the video!) the tiny move line get displayed correctly in the code display.
(of course the code modified this way cannot be used because it's incorrect).

BUT the G91 is doing something in the "code parsing process"...?

Here the video: