CubeMX UI is messed up (again, again and again...)

[Luiz Pereira]

Despite referring to the STM32F0 label (can't this forum have a simple one labeled as a general problem for STM32CubeMX???) the problem exists for ALL situations and MCU's where you simply start working with the STM32CubeMX.
Since several years ago, some colleagues on this form (and myself for 1-2 months) continue to report a completely stupid mess in the stm32CubeMX that makes it completely useless (again I attach a picture of today's nonsense and that completely prevented me simply from doing my work at the University - I'm sick of it!) No one at TS has (apparently) spent even a minute trying to solve such a stupid problem. For now (and against my will) I will migrate to NXP MCUs, which at least the configuration software works! As soon as someone at ST fixes the problem (which drags on for years!) please let me know.

[Luiz Pereira]

Hi @Paradox0v0 

Good news, you solved your problem. Unfortunately, this workaround doesn't work for me (in fact, I had already seen this workaround last year on some forum). In my case (by default the monitor is set to the recommended resolution and the font size is always 100%) this workaround has no effect: when opening stm32cubeMX, the UI problem appears after some time (which can be almost immediately or a few minutes working).

This type of problem is definitely the obvious bug of a programmer who is developing the software in a very idealistic environment, forgetting the obvious real-world application. Normally, system resources are used that, while not interfering with the programmer's system, may be restricted in other work environments.

At this point I dropped the subject. Unfortunately the problem is much more general than the STM says. When they are able to develop software worthy of the name, I will use it again.

Thanks for the info

Regards

[Herve PIERROT]

Thanks @Paradox0v0 for your feedback and workaround.

Indeed, we introduce this font setting in STM32ubeMX to workaround some issues with JavaVM not able at that time to adapt font automatically according to Window OS configuration.

Now, I understand the issue from @Luis Pereira is different as it appears after some time. We still think
it could be linked to some JavaVM settings as well, but as we are still not able to reproduce it, it's very difficult to fix.

We are still investing it anyway, and will provide update whenever we got any clue.

[Paradox0v0]

Hi @Luiz Pereira 

A few days after I solved the problem of the messing up UI, it appeared again... I encountered this problem when I tried to set the clock frequency in the "Clock Configuration" UI. If I click the selections of "PPL", the UI would be chaotic (The way to trigger the BUG has been uploaded to attachments).

Fortunately, I solved this problem again by rolling back the CubeMX to a previous version. I have installed some different versions on my desktop (v5.6.0, v6.1.0, v6.3.0, v6.5.0, v6.6.0, v6.7.0, v6.8.0, v6.9.0), and found out if installing the versions <= v6.5.0, this problem is disappeared. but if I installed the versions >= v6.6.0, this problem appeared again. By the way, JAVA on my desktop is the latest version.

[Luiz Pereira]

Hi @Paradox0v0 

Thanks for the info. In fact, I had already mentioned (although I was not specific) that the problem started to appear when STM introduced changes to the Java module. Probably it could have been from version 6.5.0 (I can try to check from the STM32CubeMX release information documentation). In any case, it is important to know whether, even in this version, the problem does not appear after some time.

In any case, downgrading is out of the question. I agree that it is a specific problem but in fact it is becoming more and more common. I reiterate my opinion that developing software (and I've been doing this for over forty years) in an "optimized" environment can only lead to failure.

When I have some free time, I will try to capture Windows events throughout the work process with the STM32Cube and then analyze them (on a time scale) and see if I have "luck" in the problem/event match.

For now, I'm still in the same situation, i.e. a few minutes after initializing the STM32Cube it becomes unusable (and yes, the problem seems to be related to user operations that involve activating the edit box controls, whatever the part of the software); on the other hand, all other software that uses Java (e.g. Matlab) has no problem.

Finally, it would be interesting to know what software STM programmers have on their PCs when developing the STM32Cube...

Regards

[Pavel A.]

> Finally, it would be interesting to know what software STM programmers have on their PCs when developing the STM32Cube...

+1. They don't have MCUXpresso, I guess))

And the same question to you @Luiz Pereira  - have you or your colleagues tried to install and run the Cube on a pristine Windows machine? 

[Luiz Pereira]

Hi @Pavel A. , @Paradox0v0 ,

There are interesting news (!)

1) Today I had to make changes in an old project (developed about two years ago) in the STM32IDE and it was necessary to reconfigure peripherals in the MCU. I opened the STM32MX (within the STM32IDE) and when I realized, I was able to work without any problems on the STM32MX for at least 30 minutes. The UI problem only appears when a popup window is presented - see figure in attachment; other than that, it works perfectly!!! (as a comparison, running the STM32MX alone, I can only work without problems in a maximum of 5 - 10 minutes...). I didn't have time to test it for long periods of work (just in case) but in the next few days I will try to do that. We will see if this new paradigm can be interesting or not.

2) Yes, I plan to test the STM32MX (alone!) on a pristine Windows machine. Just to avoid having to format a current PC (and then having to install everything again), I will use one of the new PCs that should be delivered to the laboratory next week (the software image provided by the University has only Windows 11 and Office 365). It's the "cleanest" I can "get".

[Pierre_Jean_SINACOUTY]

Hello @Luiz Pereira 

In order to push the investigation further, could you please launch CubeMX via a shell using this command of line:

java -jar -Djxbrowser.logging.level=ALL -Djxbrowser.logging.file=jxbrowser.log path/to/STM32CubeMX.exe

and then provide the generated  jxbrowser.log

BR,

Pierre-Jean

[Luiz Pereira]

Der all @Pierre_Jean_SINACOUTY , @Pavel A. , @Paradox0v0 , @Herve PIERROT 

Unfortunately, due to unavoidable commitments at the University, I did not have time to dedicate myself to the subject under analysis.
However, today I needed to use the stm32cube. I was given the option to update to version 6.10.0, which I did without any problems.
Afterwards, I opened my old project for a simple reconfiguration et voilá: when I realise, I spent more than 30 minutes without absolutely any problem in the UI interface (!!!) in any part of teh software...
I didn't do a deep test but I'll try to do it tomorrow. I note that I did not change absolutely nothing in the pre-existing settings/software + hardware drivers in my PC (I confess that I'm not sure, but I think that some update was done in in MS Office, nothing more)
If it is validated (we will see tomorrow) there was indeed a bug on the part of STM, happily (cross-fingers) fixed; and I will be happy to use STM MCUs again.

Best

Luiz

[Luiz Pereira]

Der all @Pierre_Jean_SINACOUTY  , @Pavel A. , @Paradox0v0, @Herve PIERROT 

I spoke (too) early! All it took was a new test with version 6.10 of the stm32mx for about 20 minutes and everything returned to what (unfortunately and irritatingly) was already known. The UI mess still continues..

Best

Luiz

[Luiz Pereira]

Dear all @Pierre_Jean_SINACOUTY , @Pavel A. , @Paradox0v0 , @Herve PIERROT , @mattias norlander 

I am admittedly very tired and disapointed with this issue. This is absolutely pathetic from a programming point of view and I have better things to do in my professional life than "stupidly" waste time "digging for worms".

The most obvious is a clear programming error (beginners?) adding that in Java, it is well known that there are serious problems in the graphical interface (all programmers know this!).

I note that the option "java -jar -Djxbrowser.logging.level=ALL -Djxbrowser.logging.file=jxbrowser.log path/to/STM32CubeMX.exe" in version 6.10 always gives an error (even in the administrator option) not be able to execute the command (and yes, the path to... is correctly written, pls!).

I give up. I'm sorry but I give up on STM. At this moment, and because the lab work cannot stop, I have already ordered MCU's from NXP and I am designing the PCB's for them. If anyone (by divine enlightenment!!!) at STM can understand the "blur" they are making, I would appreciate it, because I actually prefer STM's MCUs given (what I consider to be) excellent SDK work; unfortunately this is worthless for an M7 MCU where (without the stm32mx) configuration is an "impossible" job.

Just as an example of the uselessness of the stm32mx, I attach a short video (real time) of what happens when you try to run it.

https://www.dropbox.com/scl/fi/y8ksnru38ktpk106k24nu/20231207_000622.mp4?rlkey=k36it85a14fjw578t1e14vffy&dl=0

Chears