Support from ST is dead?

There's COVID-1984, and I suspect most of the key Atollic staff found other jobs.

Ultimately you get the level of support you pay for, and good support is expensive.

Ultimately you get the level of support you pay for, and good support is expensive.

Yes, you are right! But at least it would look like support! If only they would answer that they know about the mistake. But I think ignorance is a mistake.

During the quarantine I have started to learn the STM32 microcontroller because i realized that a lot industries use this kinds of microcontrollers. After two/ three months I think that, in this world, the STM32 is the worst enviroments for learn. The documentations is discutible, the forum is unusefull and the big mountain that they call cubemx enviroment, generate a lot of confusion. I have spended a lot of time, not to learn how the STM32 microcontrollers and pherihericals work, but to learn how to use the STM32 software. 

I have started with the Texas Instruments MCU's, for me the TI environment is on another planet respect STM32, unfortunately...

I think ST microcontrollers are unbeatable because of the price / value ratio and the cheap but plentiful selection of evaluation boards. I am also happy with the documentation, although for some peripherals it is sometimes too concise or incomplete / ambiguous.

What I don’t like: what they do with development environments. It was already weird that they offer two official IDEs (SW4STM32, Atollic); then these are thrown and the Atollic IDE is taken over as STM32CubeIDE.

It could have been good, but due to negligence or incompetent management, it was turned into a set of bugs.

The STM32CubeMX also contains quite a few annoying bugs that would be easy to fix, yet they haven’t done in months.

I say simple: I am a full time desktop software developer, I know what I’m talking about.

For me, microcontroller programming is a kind of outlook, though more than a hobby. I love doing it and it is sympathetic that ST has turned to open source solutions. I like that most of the software is published on GitHub. But if the CubeMX ecosystem development isn’t going that far, why not publish it as opensource software?

Information about software enhancements and updates is also a snap. There is also ST Wiki, ST Blog, but you can only find information based on words dropped from the forum.

As far as I know (although I only read on the forum) ST promised the next release of IDE and CubeMX for week 28.

I wonder if it will become something ...

I think so that the ST microcontrollers are unbeatable for the price/value ratio. I think that in the real industries this make the difference from others microcontrollers producers. But in my opinion there are a lot of problems with this cheap philosophy, first of all the discutible documentations. It's not simple to set a timer that fire at given frequency. In the datasheet the supply clock signal of the timer is called CK_INT, but who is CK_INT? If I search this word in the datasheet I not found any info about where this signal came. So the simple task, as can be a timer frequency setting, became an incubus.. This is a stupid example, but is not so stupid.. Another stupid example is the missing pinout document in the Nucleo and Discovery tools. I found the pinout of my Nucleo board in the MBED site... The question is why ST don't give me a pinout of his board?

On the software side, I'm one that like the model based philosophy, so I agree with the code generator tools. In this case, in my opinion, the ST tools are too rigid and too poor to be used. There a lot of confusion about IDE, firmware and tecnology. It's very hard for me to find a work enviroments that i can use to develop some code. The CubeIDE it's nice, but the example are for others IDE, also the TouchGFX generate a project structure that it's different from one that is generate from the CubeIDE. Obviosly there board support missing, so if you want to use a pin on the Nucleo board, you need first of all to find a pinout of the board, and after set the pin on CubeMx editor..

In this days I tred to use the STemWin framework.. In the all documentations that I found, is shown that by using the CubeMX, the STemWin project is trivial to setting. The problem is that the newest version of CubeMX don't support STemWin.. I'm not able to find any documents that explain how set-up a StemWin project using the new version of ST software tool.. At this moment, i use a old version of CubeMx (5.3) to generate a project for STemWin..

Yesterday I have create a post where I ask how to use the firmware examples, because i'm not able to use the examples and modify. Seem that this task it's not trivial as you can expected..

This it's a big picture on what i do in this last two months.. Honestly, i'm not very happy..

Since I’m a desktop c ++ developer with some minimal and old mcu knowledge (mostly 8-bit 8051), it was a slow learning curve with a lot of confusion and contradiction in the beginning.

I think with no previous ARM/Cortex-M knowledge You haven't any chance to achieve any result with only the ST's documentation.

I started using STM32 MCUs with two must-to-have books:

Embedded Systems with Arm Cortex-M Microcontrollers in Assembly Language and C : Third Edition by Yifeng Zhu

and

The Definitive Guide to ARM (R) Cortex (R)-M3 and Cortex (R)-M4 Processors by Joseph Yiu

About the pinouts of the boards:

All information (including circuit diagrams) on the Nucleo-144 boards can be found in UM1974 (User manual ofSTM32 Nucleo-144 boards).

I feel ST is at fault in that the documents are published in bulk on the product page.

The beginner has no idea in which documents you will find vital information.

I also started like you: google my friend ...

I have the user manual, but it's in ST style, in my opinion is much better to have this type of document for the pinout.

I have read both the books that you suggest, they are very useful for learn. The problem however isn't the Arm Cortex architecture, the problem is the ST software! I have spended a lot of time in this two/ three months only to understand how to correctly use the ST software and I still have a lot of problems.. This is why i'm not happy..

I can't find a clear way on how to use their software. At the moment that I think that I have understand something about the ST software, everything I understood turns out to be useless.. Think about the STemWin framework.. At this moment is not possible, or is better i'm not able!, create a project using the CubeMX or the STM32CubeIDE that generate the STemWin code. You need to ad manually the STemWin framework and the LCD drivers. This is in contracdition with the ST software philosophy. So I really need to learn the ST software?

I have some doubts about this.. Maybe tomorrow, they change something and all that I have learned became unuseful..

Fortunately exists google.. ;)

When you start, i recommend use Keil , or cube with export project to Keil.

Keil IDE it's a nice tool, but it's not free. The trial version have a size code limit if I remeber..

KEIL MDK and IAR EWARM are proprietary tools for Windows only for horror prices.

Okay, professional tools with lots of useful features. But the world is changing and more and more companies are turning to the open source world.

I have tried both IDEs and the debug features as well as the included middleware (+ source code) are really valuable features.

Professional work requires professional tools, and large companies can pay for it.

BUT!

The code editing features of both development systems reflect the era of Windows 95-98. It is light years behind even an Eclipse or VSCode.

Many argue that proprietary compilers generate much better code than the GNU or CLang toolchain. In the case of compilers, I cannot comment because I do not have serious experience or competence. I think free compilers are evolving, and ARM itself is developing the GNU toolchain.

I basically love the STM32CubeIDE as much as I loved its predecessor - Atollic.

That’s why I’m worried about the future of the software: if what Clive1 claims is true (there’s virtually no serious developer behind the project), we can’t expect much good.

For me, this is confirmed by ST’s striking ignorance of bugs affecting the IDE.