What’s new in STM32CubeIDE 2.0.0

I'm a recent user of CubeIDE / MX and I have to comment just how appalling the new MX to IDE experience is. 

OK so I have to create an .ioc file in a separate MX package - no problem.

But why when I try to load that ioc file (or import or build a new project around it) is it such a comically convoluted process?

Is it too much to ask that 'import' actually lets me see the ioc file and import it?  Rather than some random finder window that apparently can't see ioc files - which seems like quite a big requirement to miss?

So maybe I should be setting up the project ... in the MX ?  When I'm actually going to be developing in the IDE?  Well OK that makes zero sense but let's give it a go ... wait, the project manager forces the new project to generate under root (even if you unclick it) and greys out the project location, then gives you a warning it will nuke the ioc you just developed if you so much as dare try to actually start the project you were only trying to start in MX because IDE apparently forces you to?

What's bizarre is I was using the previous version and it was a smooth experience.  All this new friction and weirdness and I'm sure it must be worth something to make the dev team look busy for their annual appraisal, but as a User Experience is is either (a) incredibly badly thought through or just (b) uses unintelligible and un-intuitive menus which make even the simplest task ('opening an ioc file in a new project') so much more complex and time consuming than it needs to be.

Maybe there are some power users who enjoy learning arcane and convoluted workflows, but for non-masochists, this is not a good piece of software.

Why create these wizards for power users when I assume power users are the last people who need or want wizards?  Why make people who aren't power users now have to try to determine which of the 20 import options hidden in 6 folders is the one that they need?   Surely if, say, your software is importing an 'Atollic TrueSTUDIO project' you could get your software to look at the file format rather than making users wade through these lists?  Surely importing from 'File System' under 'General' is already covered by C++ or vice versa?  So why introduce so many new ways to achieve the same thing when all of these new ways could just have been handled by getting your software to look at a file format? 

(I'll take all this back if someone can explain how to open an ioc file created in Mx, in  CubeIDE, in three mouse clicks or less)

Sad to see STM joining the race to the bottom to move hardware IDEs to VS Code. It has recently been done by many big companies like TI and Microchip and their implementations are so buggy that they are almost unusable. Just like many features were broken in this new update, how many features will be broken to accommodate this jump to VS Code? Things like @Orion33 pointed out such as debugging floating point numbers or even being able to plot using Serial Wire Viewer just to name a few. It seems to me that this decision has little to to do with memory savings, updatability, and whatever else was stated and is more the result of perceived market forces such as competitor decisions and the push for AI integration in everything. Rest in peace STM32CubeIDE.

A compromise?

Putting MX back into Cube would restore order and make a lot of people happy...

Do MX releases as often as is needed.

Update Cube occasionally.

This way if you want to use what is already supported in Cube then you are happy.

If you need to use a new CPU then use MX

Simple!

Forget VS Code, it is like going to the 80's

Don't even think about AI ;)

@sksouri

I agree. If somebody told you to put your hand in a fire would you?

How this works with the user code sections?

Can I try the code, make some changes change in cubeMX and regenerate the code without loosing the user code?

@coso2 coso2 

If it is in an MX generated file and NOT within the USER CODE BEGIN and END you will lose it on re-generation.

I've known about the VisualGDB plugin for a long time, but only decided to try it now: https://visualgdb.com/?features=embedded

Surprisingly, these third-party developers even have live variables that work without constantly opening random code files (who knows, this is an old CubeIDE problem that also occurs in VSCode + STM Plugin), plus they can be displayed immediately as graphs without distracting you with writing the backend — an amazing solution! The plugin also works with CubeMX, so I see no reason to use CubeIDE + CubeMX instead of this combination.

Can anyone point me at a step by step guide.  I can generate code using CubeMX,  Then how do import this into CubeIDE.  When I do there is no builder setup? I just need a quick start guide !!   

EDIT:  Found it!   In CubeMX, project settings, Set the tool chain to to STM32Cube_IDE. 

Hello, 

@TimFrancis, please find below the different steps and let me know if you need further details:

Case1: Create new project

1. Open STM32CubeMX and set your configurations
2. Go to “Project Manager”
3. Add project name and select STM32CubeIDE
4. Do “GENERATE CODE”
5. Open STM32CubeIDE V2.0.0
6. Import your project: put your folder path where the project is created
7. The project will appear in the STM32CubeIDE workspace
8. Compile your project

 Case2: Update project

1. You double clic on .ioc file in STM32CubeIDE workspace
2. MX project is opened
3. Add your updates
4. Redo steps of Case1: from 2 to 4
5. Go back to STM32CubeIDE
6. Right clic on your project folder in workspace and do refresh (F5)
7. Compile your project

please do a video tutorial, for new stm32cube ide 2.0.0,

Yaswanth,

see lbthomsen's post above. Here is the direct link to the video, https://www.youtube.com/watch?v=Sa_HBrblF0w