STM32CubeIDE for Visual Studio Code (MCUs)

Hi,I’m trying to understand how the build setup works in VS Code.Could you advise where and how to configure VS Code so that .bin and .hex files are always generated when building both Debug and Release configurations?Is this controlled via CMake, or through one of the .json configuration files?Whichever approach is correct, could you also share the specific scripts or commands that need to be used?Thank you :) 

Hi all,What is the recommended way to not crush our repository and dynamically pull/generate the project's Drivers and Middleware folders?  For instance, our repository has both Bootloader + Application code projects in a repo (and other MCU projects/layouts for the same board).Is there something similar to Platform.IO (or C#'s NuGet packages) where the platform maintains those 3rd-party files in a volatile location and pulls them if they don't exist on your machine? Especially for the sake of not checking into the repo duplication of +15 MB worth of files for each project.Please excuse me if this was already asked. I didn't quite find a solid answer.I'm ok with making a command line script that auto-pulls them too, if that's a possibility.

Hello, Long time lurker here. First of all, thank you for the extensions, really appreciated the possibility to work directly in vscode.I have a couple of situation, however, that are really annoying and that I would like to solve. While developing in vscode with the extension enabled. It is sufficient to update the CMake presets in order to have the tools installed by the ST extension to make the git repository dirty.If before committing the files to the repo a pre-commit that formats the files in a proper way then it is amplificated.The files: c_cpp_properties.json in the .vscode folder is always updated with the compileCommands updated if dynamic folders for Debug and Release are used .clangd configuration file is updated in the CompilationDatabase field.In my opinion there are two different types of settings: settings that needs to be persinstent, and maybe reusable on different computers by different userssettings related only to the current runThere

Hi, I consistently hit HardFault when using PICOLIBC. I am using a STM32H735G-DK board.Setup:1. Start from a black CubeMX project, selecting the board and default pins. The FreeRTOS must be enabled, and the TouchGFX (with the MPU configured). We will target CMake and SR Arm Clang. Generate code.2. Open TouchGFX and load the project with the file ApplicationTemplate.touchgfx.part. No need to add anything, just generate code.3. Open the entire project with VSCode. The project by default uses the PICOLIBC in the file starm-clang.cmake. The project as is compiles successfully. flash onto the board and click Continue. The code will instantly go in HardFault_Handler().Here is a printscreen: Can anyone try to reproduce this, just to make sure it is not something only on my side? Thank you!If I change for STARM_NEWLIB in starm-clang.cmake instead of STARM_PICOLIBC, everything works as expected, no hardfault, and the screen initializes successfully.I have added a demo project.

Disclaimer:Probably not quite the right place for this bug report. However, since the issue appeared with the update to CubeIDE VSCode v3.8.0 and STM32CubeMX v6.17.0, and the project in the attached ZIP was created with these versions to reproduce the behavior, I am placing it here.IntroductionLink to Project on GitHub: init_array_issue (or use the attached ZIP file)The code in file init_array_issue.cpp demonstrates a misbehavior where, under certain conditions, the initialization of objects is silently skipped without any warning or error message when picolibc is used together with the linker description file STM32H753XX_FLASH.ld generated by STM32CubeMX and the default optimization -Og.picolibc / STARM_PICOLIBCThe use of picolibc has been the default since the STM32CubeMX update to 6.17.0, configured via the option STARM_PICOLIBC in the file .\cmake\starm-clang.cmake.Previously, STARM_HYBRID was set here by default.# STARM_TOOLCHAIN_CONFIG

Hello,Before compilation:   I’m having an issue with IntelliSense (possibly clangd).When a macro is undefined, the code inside #ifdef is correctly shown as inactive (dimmed). However, after I build and flash the project, the same code becomes highlighted as if it were active, even though the macro mDEBUG is still not defined.The highlighting only returns to the correct state after I switch tabs or force a refresh.Is this expected behavior, or could this be a bug or bad config on my side? It brings confusion to my workflow.After compilation: 

Quick question. I am working on two projects:One is a dual-core project, and uses the STM32H755 and CMSIS-RTOS v2 over Free RTOS. I often have to debug two boards simultaneously.Another project is using STM32H573. This uses TrustZone and ThreadX (on the non-secure side).Both are currently managed via ST Cube IDE (Eclipse based of course) and debugging is already quite "complex".I can see advantages with VS Code, but moreover, I would not be surprised if the VS Code route becomes the de-facto standard some day (purely speculative of course), so I think I may need to start looking at it.Before I do, my question to the community is whether the VS Code option is yet mature enough for dual core and TrustZone projects?Many thanks in advance,Nick

I have same error since days on differnt machines in different networks.While this is working:cube bundle list-online PS C:\git\nucleo-h753zi-canfd> cube bundle list-online Name : Platform Version --------------------------------------------------------------------- adoptium-jre : x86_64-windows ["21.0.8+9.st.2", "17.0.16+8.st.1", "17.0.15+6.st.1", "17.0.14+7.st.6", "17.0.14+7.st.5"] cmake : x86_64-windows ["4.2.3+st.1", "4.0.1+st.3"] cube-cmsis-scanner : x86_64-windows ["2.4.2", "2.2.1", "1.1.3", "1.1.0", "1.0.1"] cube-code-doc : all ["0.0.7+st.2", "0.0.6+st.2", "0.0.5+st.1", "0.0.3+st.1", "0.0.2+st.1"] cube-code-manifest : x86_64-windows ["1.1.3+st.1", "1.0.9+st.2", "1.0.6+st.1", "1.0.3+st.1", "1.0.1+st.1", "1.0.0+st.1"] cube-wrapper : x86_64-windows ["0.10.2", "0.8.3", "0.7.3"] gnu-gdb-for-stm32 : x86_64-windows ["14.3.1+st.2", "13.3.1+st.10", "13.3.1+st.9"] gnu-tools-for-stm32 : x86_64-windows ["14.3.1+st.2", "13.3.1+st.9", "13.3.1+st.8"] gnu-tools-for-stm32-13_3_1-d

My STM32cube IDE project has a pre-build command:${ProjDirPath}\Middlewares\Third_Party\nanopb\generator\protoc --nanopb_out=${ProjDirPath}\Application\cm4_interface --proto_path=${ProjDirPath}\Application\cm4_interface ${ProjDirPath}\Application\cm4_interface\cm4-interface.protoThe json that's generated doesn't correctly escape the '\' characters, which results in an error...   

.clangd is constantly being modified appending the same compileflags: CompileFlags: Add: - '-ferror-limit=0' - '-Wno-implicit-int' CompilationDatabase: CM7/build Diagnostics: Suppress: - unused-includes - unknown_typename - unknown_typename_suggest - typename_requires_specqual If: PathMatch: CM7/.* --- CompileFlags: Add: - '-ferror-limit=0' - '-Wno-implicit-int' CompilationDatabase: CM4/build Diagnostics: Suppress: - unused-includes - unknown_typename - unknown_typename_suggest - typename_requires_specqual If: PathMatch: CM4/.* --- CompileFlags: Add: - '-ferror-limit=0' - '-Wno-implicit-int' CompilationDatabase: CM4/build Diagnostics: Suppress: - unused-includes - unknown_typename - unknown_typename_suggest - typename_requires_specqual If: PathMatch: CM4/.* --- CompileFlags: Add: - '-ferror-limit=0' - '-Wno-implicit-int' CompilationDatabase: CM7/build Diagnostics: Suppress: - unused-includes - unknown_typename - unknown_typename_suggest

 reason: Error: Failed to get full list of project bundles to install. Command failed w Steps I've already triedWindows11Verified the project path contains no Chinese characters, spaces, or special symbols.Cleaned and rebuilt the project (CMake/Make workflow).Reloaded the VSCode window Confirmed the executable file was generated successfully, and the program path in launch.json points to the correct file. QuestionsWhat exactly causes this "Failed to list bundles" error during debugging?Is there a standard, correct launch.json template for this scenario on Windows?Do I need to configure additional GDB paths or environment variables?Are there any other troubleshooting steps I should try? Thanks a lot for any help or suggestions! :folded_hands:

Hi, I am not sure how to categorize this problem. I have a custom board that I had done an initial software development cycle on STM32CubeIDE. I am working a newer version in VSCode with the STM32 Extension. Everything is working with the exception of one problem. I am using a watchdog on my MCU (STM32L4R9ZGJ6), which triggers while I have the processor halted. I know how to fix this in CubeIDE (Debugger tab, Enable Suspend Watchdogs) but cannot figure it out on the VSCode/GDB side. Below is the launch.json I have so far:{ "version": "0.2.0", "configurations": [ { "cwd": "${workspaceFolder}", "executable": "./build/debug/build/test.elf", "name": "Debug with ST-Link", "request": "launch", "type": "cortex-debug", "runToEntryPoint": "main", "showDevDebugOutput": "none", "servertype": "stlink" }, { "name": "Launch", "type": "cppdbg", "request": "launch", "cwd": "${workspaceFolder}", "program": "${command:cmake.launchTargetPath}", "MIMode": "gdb", "miDebuggerPa

why?

Hello,With the new STM32Cube Debug Jlink GDB Server V1.2.0 I'm not anymore able to start the debugging session. If I downgrade to V1.1.0 then it is working again. I get always the error message "No GDB Server running". I tried that with my old launch.json settings and with the new generated and adapted json settings.Do I miss some special settings or has the new extension a bug?Or do i need any dependency? error message: old launch.json new launch.json  Thanks in advance,nobody 

Hi everyone,I found a solution for a common issue when using the official STM32Cube for VS Code extension. If you disable the Microsoft C/C++ (ms-vscode.cpptools) extension to avoid IntelliSense conflicts, you might notice that "Format Document" stops working completely.The reason is that the STM32 extension's built-in clangd has its formatting engine disabled by default.To fix this:Go to VS Code Settings (Ctrl + ,).Search for: stm32cube-ide-clangd.enableFormatting.Check the box to enable it.If prompted, set "STM32Cube for VS Code" as your Default Formatter.This allows you to have working formatting and precise IntelliSense without any conflicts from the Microsoft extension. Hope this helps! 

Hi,I am using the latest version of STM32Cube extension for VS Code (3.8.0). I created a simple project for STM32H735G-DK board using CMake as the toolchain and ST-ARM-CLANG as the compiler. I do not know if it is relevant, but I also enabled FreeRTOS and configured the USE_NEWLIB_REENTRANT option to Enabled.I generated the project and I am compiling it with VSCode. I changed in the starm-clang.make to use the NEWLIB instead of the PICOLIBC:set(STARM_TOOLCHAIN_CONFIG "STARM_NEWLIB")The compilation works, but clangd complains about some errors. In the output of clangd, I can see multiple instances of:D:\ST\stm32cube\bundles\st-arm-clang\21.1.1+st.6\bin\..\lib\clang-runtimes\multilib.yaml:47:3: error: unknown key 'IncludeDirs' IncludeDirs: ^~~~~~~~~~~ D:\ST\stm32cube\bundles\st-arm-clang\21.1.1+st.6\bin\..\lib\clang-runtimes\multilib.yaml:21:1: error: multilib "arm-none-eabi/armv4t_exn_rtti_size" specifies undefined group name "stdlibs" MultilibVersion: '1.0' ^and because of these erro

Hi,After updating to the latest VS Code extension pack v3.8, what is the recommended path for executing CMake commands from command line? I uninstalled the STM32CubeCLT since it's now bundled, perhaps this the wrong choice?I want to test running commands to upgrade our CI/CD pipelines to be up-to-date.  Also found following path which contains `cube.exe` as a CLI wrapper. My concern is setting an environment variable to a path with a version number (`/...-core-1.2.0-.../`) that ends up being a moving target.> `C:\Users\USERNAME\.vscode\extensions\stmicroelectronics.stm32cube-ide-core-1.2.0-win32-x64\resources\binaries\win32\x86_64`Any assistance would be greatly appreciated. 

After a recent update I am unable to run my code from vscode on my Windows 11 machine. The compilation succeed, but when I click on Run -> Start Debugging nothing happens. The System Bundle Manager shows that I have installed all the latest versions, but the bundle manager of the projects shows only a few of those bundles (see attached). How can I install the bundles in the project section?Note that I have already tried the usual things: remove the build, .vscode and .settings folders, rebooting the pc, upgrading the stlink firmware.I tested the same code on another linux pc and there it works as expected. On the linux PC `stlink-gdbserver` was installed at project level

After upgrading STM32CubeIDE for Visual Studio Code to version 3.8.0, the code formatting function is no longer available. Currently, while writing the code and entering ";", the formatting feature is not functioning. When saving the code, the code will not be automatically formatted; both "Format on Save" and "Format On Type" in the settings have been checked.

I'd like to clone an existing STM32 C project with CMake that I've previously setup. I've seen posts about how to do this on STM32CubeIDE but none using the VSCode extension.Currently, I'm aware of 2 methods:Cloning `.ioc` fileClone the `.ioc` file into new folder with <New Project Name>.Find and replace any <Old Project Name> references with <New Project Name> in the `.ioc` file.Open the updated `.ioc` file in CubeMX and generate code.Copy over any user code from the original `main.c`.(a bit tedious depending on how much user code there is)Copy over any libraries and recreate any symlinks.Build projectCopying the whole project folder and renamingCopy entire project folder and rename it.Find and replace all occurrences of <Old Project Name> with <New Project Name>.Try building and fix any errors that pop up.Method 1 feels safer but is more tedious than method 2.Is there a neater way to do this that doesn't involve manually replacing values? (i.e. with

IntroWhen I'm using the "STM32CubeIDE for Visual Studio Code" extension, I often want to edit C source files outside of my workspace, which are listed in my CMakeLists.txt and get compiled into the firmware just like any other C file.  Unfortunately, with the default setup, the language server (clangd) cannot find some of the headers I include, and it doesn't know the right compilation options, so intelligent features like inlay hints (red squiqqlies) and the "Go to definition" command do not work properly.  The ST extension attempts to communicate this kind of information to the language server by putting a ".clangd" file at the root of my workspace folder which has a line in it like "CompilationDatabase: build_release".  But clangd won't consider that file when looking at source files outside of the workspace.MotivationI have multiple STM32 projects that share some common libraries, so my folder structure looks like:lib/lib1/lib1.clib/lib1/lib1.hlib/lib2/lib2.clib

Hi, I am seeking some guidance.With the new release STM32CubeIDE for Visual Studio Code 3.8.0, it says it comes with new features including " Serial terminal: Enabling developers to easily setup serial communication for example leveraging the ST-LINK virtual com port."In the docs, there is not much explanation on how to get this working. I only see:"If the serverPort attribute is defined, two ports are used by the ST-LINK GDB server: one for the debug session and serverPort + 1 for the SWO connection.These ports are reserved even if SWO is not activated in the debug configuration."How would one activate the SWO in the launch.json? I went through all the options with completition and none seem to activate or configure the SWO.I know that with the extension "Cortex-Debug" provided by marus25 we could configure the SWO as so:"swoConfig": { "enabled": true, "cpuFrequency": 275000000, // must match -a (Hz) :contentReference[oaicite:4]{index=4} "swoFrequency"

Having some issues trying to debug a FreeRTOS project using JLink, using an STM32U5G9 (Riverdi 5" Display Board) If I enable "serverRtos" in the launch configuration:"configurations": [ { "type": "jlinkgdbtarget", "request": "launch", "name": "STM32Cube: STM32 Launch JLink GDB Server", "origin": "snippet", "cwd": "${workspaceFolder}", "preBuild": "${command:st-stm32-ide-debug-launch.build}", "runEntry": "main", "serverSpeed": "12000", "serverRtos": { "enabled": true, "port": "60000", "driver": "freertos", "driverPort": "ARM_CM33_NTZ" }, "preRunCommands": [ "monitor exec SetRTTSearchRanges 0x200bb800 0x234800" ], "imagesAndSymbols": [ { "imageFileName": "${command:st-stm32-ide-debug-launch.get-projects-binary-from-context1}" } ] } ]And then try to start a debug session, i get a disconnect error:Reading 20 bytes @ address 0x200C2A5C Read 4 bytes @ address 0x200C2A28 (Data = 0x200C2970) Read 4 bytes @ address 0x200C2A58 (Data = 0x200BEA20) Read 4 bytes @ add

Hi ST-Team,context:- STM32H753ZI NUCLEO Board- VSCode with STM32CubeIDE Extension- Clang Compilerafter follow updatesSTM32CubeIDE VSCode from v3.7.0 to v3.8.0STM32CubeMx from v6.16.1 to v6.17.0incl. MCU Package H7 from v1.12.1 to v1.13.0incl. Azure RTOS for H7 from v3.4.0 to v3.5.0we get now following warning if a file is including the header\Drivers\CMSIS\Device\ST\STM32H7xx\Include\system_stm32h7xx.h[build] [6/206] Building C object CMakeFiles/STM32_NUCLEO-H753ZI.dir/AZURE_RTOS/App/app_azure_rtos.c.obj [build] In file included from D:/dev/Prj/Clean_STM32H7_Project__02/AZURE_RTOS/App/app_azure_rtos.c:28: [build] In file included from D:/dev/Prj/Clean_STM32H7_Project__02/cmake/stm32cubemx/../../Drivers/CMSIS/Device/ST/STM32H7xx/Include/stm32h7xx.h:124: [build] In file included from D:/dev/Prj/Clean_STM32H7_Project__02/cmake/stm32cubemx/../../Drivers/CMSIS/Device/ST/STM32H7xx/Include/stm32h753xx.h:232: [build] D:/dev/Prj/Clean_STM32H7_Project__02/cmake/stm32cubemx/../../Drivers/CMSIS/De

 stm32cubeIDE for Vscode ，there is no one-click download function. How to add?