Forum Replies Created
-
Posts
-
I have got my base project to compiler and link. I had to add one more linker option “-nostartfiles”, this fixed the above error.
Now I will proceed to create a template from this project and then rebuild my real project using that template.
Regards,
Sid
Thanks, I forget that many options may be set in the MS Build Properties dialog.
I have my custom linker script installed and now I am getting a linker error that I do not understand. Any pointers would be much appreciated. I have attached the verbose output log.
Sid
I am trying and failing to get a non-standalone project to do what I need. I am doing this so that I can then generate a custom template for my projects. These projects need to use libOpenCM3 and this is not an option in VGDB.
First I created a new embedded project for STM32F401RE, choosing the simple HAL-Blink example.
Then, because I do not want to use HAL I need to use libOpenCM3, I removed that package from the project using the VGDB properties dialog.
I also selected the Newlib-nano with floating point support in printf (required) and hardware floating point in the VGDB properties dialog.
No semihosting was also selected.
Since libOpenCM3 library brings the startup file I deleted startup_stm32f401rexe.c from the project.
I also need the project to use my linker script, however, I see no way of doing this in the VGDB properties dialogs, the default linker script is in the file “stm32.props”, however there is a warning there to not edit that file.
So, how can I get my script used?
Sid
My code needs to use printf through a USB-UART for a configuration interface for the device.
It looks like I really have no option but to create a new project, this is so disappointing and something the project schedule could do without.
Sid
I created the project from a custom template and do not think I converted it to standalone, perhaps the template was standalone. It took a long time to get this project building correctly. The custom template was created in order to have a template for libOpenCM3.
Is there no way to change the project options to turn semihosting off?
Sid
That would be a huge job, it is a large project and would take a long time. Is there no other way to do this.
Not being able to easily migrate projects when there are toolchain updates and is a poor user experience,
Sid
Thanks for the reply, I don’t have mention of rdimon.specs in my LDFLAGS:
-Wl,–defsym,_stack=0x20006000 -Wl,-T,src/platforms/stm32/f4discovery.ld -nostartfiles -lc -lnosys -Wl,-Map=mapfile -mthumb -mcpu=cortex-m4 -Wl,-gc-sections -mfloat-abi=hard -mfpu=fpv4-sp-d16
So, I am a bit lost with your instructions,
Sid
Thanks, the options are not available in my project. However, on the “Embedded Debug Tweaking” option page I see the enable/disable setting.
Sid
I would accept the explanation if the red box could be removed without having to close the file, that does not seem like proper behavior and it counter-productive when one is writing code. Clicking on the “light bulb” options does not remove it.
Sid
Thanks for the suggestion but that cannot be the problem because it happens when I type a function for the first time. As I said if I make a mistake typing the function name, then go back to correct it the red box appears and will not go away until I close the file and reopen it. At the time there is no header file that has the function name.
Easy to reproduce:
- type “void aFuntion(void)
- Before hitting enter go back and correct the function name to be “aFunction”
- go to end of the line and hit enter
- Now you have red box until closing and reopening the file.
Sid
Sid
Thanks, it is annoying but not a show stopper. I would be happy to work with you to trace the issue if you would like,
Sid
