Hi,
Maybe I was jumping in to conclusion and the issue is not the same as
what I had. I was building PX4, which uses CMake build system, so I am
not having any Makefile or Make.defs in my own board directory. Also the
platform is not stm or arm, but risc-v.
Anyhow, this is the error which I
> It seems like Daniel is hitting the same issue
Daniel is actually not using the common folder from the STM32
directory. This is why he had to do that renaming.
The issue is this:
https://github.com/apache/incubator-nuttx/blob/master/tools/Config.mk#L154-L156
Which is forcing a common directory
Hello Abdelatif,
"Which is forcing a common directory when there isn't one. " -- This
statement is not true as we have
https://github.com/apache/incubator-nuttx/blob/b953296de73ac75bb380a609f9f30e9fe34e7622/tools/Unix.mk#L272-L277
that depends on if "BOARD_COMMON_DIR" exists or not.
Getting back
> "Which is forcing a common directory when there isn't one. " -- This
> statement is not true as we have
> https://github.com/apache/incubator-nuttx/blob/b953296de73ac75bb380a609f9f30e9fe34e7622/tools/Unix.mk#L272-L277
> that depends on if "BOARD_COMMON_DIR" exists or not.
Unix.mk includes $(TOPD
> I think with NuttX 10.2.0 you do not need to perform the next steps any more.
These are necessary when someone is using a custom board without a
common folder copied from a board in-tree that contains a common
folder.
On Thu, Mar 3, 2022 at 11:31 AM Abdelatif Guettouche
wrote:
>
> > "Which is
If what you are writing is true, then compilation for any board that
does not have a "common" folder at board level should fail. And there
are plenty of boards in NuttX tree that compile without an error (CI
on mainline pass without any problems). I'm not saying that
"BOARD_COMMON_DIR" will not hav
Hi,
There is no problem with that. We just need to copy source files from
NuttX board common to custom board location and add files to
compilation list together with disabling of NuttX board common layer
(set CONFIG_BOARD_STM32_COMMON=n for example with STM32 case).
Or maybe I'm missing the exact
> If what you are writing is true, then compilation for any board that
> does not have a "common" folder at board level should fail.
No because in-tree this line: BOARD_COMMON_DIR = $(wildcard
$(TOPDIR)$(DELIM)boards$(DELIM)$(CONFIG_ARCH)$(DELIM)$(CONFIG_ARCH_CHIP)$(DELIM)common)
will result on an
Hi Jukka,
I tried to replicate your case, but failed with:
$ git submodule update --init --recursive
Cloning into
'/home/pkarashchenko/workspace/px4-firmware/boards/ssrc/saluki-v1'...
ERROR: Repository not found.
fatal: Could not read from remote repository.
Please make sure you have the correct
Hi, sorry about that; you can just remove that submodule, it is not needed to
re-produce the issue.
- Jukka
Petro Karashchenko kirjoitti torstai 3. maaliskuuta 2022:
> Hi Jukka,
>
> I tried to replicate your case, but failed with:
> $ git submodule update --init --recursive
> Cloning into
> '/
And also note that the commit I mentioned is already reverted in nuttx
submodule, so you need to put it back in order to re-produce the issue :)
Jukka Laitinen kirjoitti torstai 3. maaliskuuta 2022:
> Hi, sorry about that; you can just remove that submodule, it is not needed to
> re-produce the
Hello Abdelatif,
I just tried to recreate the same scenario. To do this I did:
1. Recreated folder structure as in Daniel's case
|-> apps
|-> my-folder
|-> my-boards
|-> custom-board
|-> nuttx
2. Copied nucleo-g431kb to custom-board: cd my-folder/my-boards && cp
-r ../../nuttx/boards
> Compilation failed, but not because of a common folder but because:
Again, the issue is to not use the common folder at all. This was
requested before. What you did here is that you used the one in-tree.
> I think this is a weak argument as we change config
option names sometimes as users are
The case is only about creating a proper symlink. "Again, the issue is
to not use the common folder at all." -- agin, that is possible. Each
"board/common" has options like "BOARD_STM32_COMMON" or
"BOARD_SAMV7_COMMON" and that is exactly to or not to use in-tree
"board/common". What other option do
> The case is only about creating a proper symlink.
I think in the current situation the symlink is created unconditionally.
> Abdelatif, I would really appreciate it if you can write and send me
some small proposal at what level the option should exist.
What I was referring to is similar to opt
Hi,
"I think in the current situation the symlink is created
unconditionally." -- yes. If the in-tree board has a common folder the
symlink points unconditionally to that location.
Introducing "CONFIG_ARCH_BOARD_COMMON" may be a good option here, so
the user can explicitly specify to create a sym
Hello Jukka,
Do you have any steps on how to build or download a docker image that
can be used to build that project?
It takes me too much time to get build running.
Best regards,
Petro
чт, 3 бер. 2022 р. о 12:43 Jukka Laitinen пише:
>
> And also note that the commit I mentioned is already reve
Hello Petro,
Did you use any specific feature from boost?
Most used features from boost were integrated to newer versions of C++.
I used C++-17 in a project and it worked fine.
Best regards,
Flavio
Em qua., 2 de mar. de 2022 às 11:01, Xiang Xiao
escreveu:
>
> We use both uClibc++ and libc++,
Hi, sorry but I am away from computer atm.
In the project there is "packaging" subdirectory, it contains some docker env,
which is also used in TII's CI
- Jukka
Petro Karashchenko kirjoitti torstai 3. maaliskuuta 2022:
> Hello Jukka,
>
> Do you have any steps on how to build or download a dock
Hi Alan,
Thanks for the participation!
In the end I've dug into the driver code and found the bug with CAN3 filter
configuration.
Here is the fix: https://github.com/apache/incubator-nuttx/pull/5677
---
With best regards, Oleg.
чт, 24 февр. 2022 г. в 23:01, Alan Carvalho de Assis :
> Hi Oleg,
Hi Petro,
I forgot to mention that I need to enable BOARD_CUSTOM_LEDS, sorry about
that. Now I was able to build without errors keeping src/Make.defs
untouched.
Thanks everybody
Daniel Pereira de Carvalho
Em qui., 3 de mar. de 2022 às 09:20, Petro Karashchenko <
petro.karashche...@gmail.com>
Hi Oleg,
I'm glad you figured it out.
BR,
Alan
On 3/3/22, Oleg wrote:
> Hi Alan,
>
> Thanks for the participation!
>
> In the end I've dug into the driver code and found the bug with CAN3 filter
> configuration.
> Here is the fix: https://github.com/apache/incubator-nuttx/pull/5677
>
> ---
> W
Hi
where can I found a simple example that send n bytes to the SPI and
receive m bytes?
I can now see the /dev/spi but using simple methods like open and write
doesn't give me any signal on the bus...
Thanks in advance
Roberto
23 matches
Mail list logo
