Re: Cygwin tools failure

[Joel Sherrill]

On Thu, Oct 13, 2022 at 5:32 PM Ryan Long  wrote:

>
> On 10/12/2022 7:43 PM, Chris Johns wrote:
> > On 12/10/2022 3:24 pm, Chris Johns wrote:
> >> Consider a patch to add the piece removed at the end of this patch as
> approved.
> > I see the patch was pushed, thanks. The RSB hash will need to be updated
> so
> > users pick it up.
> I've sent a patch that updates the hash to Joel. He'll get that pushed
> soon.
>

And pushed.

--joel

> >
> > Thanks
> > Chris
>
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: Cygwin tools failure

[Ryan Long]

On 10/12/2022 7:43 PM, Chris Johns wrote:

On 12/10/2022 3:24 pm, Chris Johns wrote:

Consider a patch to add the piece removed at the end of this patch as approved.

I see the patch was pushed, thanks. The RSB hash will need to be updated so
users pick it up.

I've sent a patch that updates the hash to Joel. He'll get that pushed soon.


Thanks
Chris

___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: Cygwin tools failure

[Chris Johns]

On 12/10/2022 3:24 pm, Chris Johns wrote:
> Consider a patch to add the piece removed at the end of this patch as 
> approved.

I see the patch was pushed, thanks. The RSB hash will need to be updated so
users pick it up.

Thanks
Chris
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: Cygwin tools failure

[Chris Johns]

On 12/10/2022 2:30 pm, Joel Sherrill wrote:
> 
> 
> On Tue, Oct 11, 2022 at 8:36 PM Ryan Long  > wrote:
> 
> Cygwin has been unable to build the tools for awhile. This has been due
> to an array subscript being a char while building DTC. The maintainers
> didn't see an issue with the code itself, so it's going to go on being
> broken for awhile I guess.
> 
> Joel recommended commented out the building of DTC through the RSB.
> After doing so, I'm running into the following error. I was building
> AArch64's tools. Disabling DTC and building with my Debian WSL instance
> built it just fine.
> 
> 
> [ 48/258] Compiling rtemstoolkit/elftoolchain/libelf/libelf_align.c
> ../rtemstoolkit/elftoolchain/libelf/elf.c:34:35: error: ‘LIBELF_ARCH’
> undeclared here (not in a function); did you mean ‘LIBELF_ERROR’?
>     34 | .libelf_arch    = LIBELF_ARCH,
>    |   ^~~
>    |   LIBELF_ERROR
> ../rtemstoolkit/elftoolchain/libelf/elf.c:35:35: error:
> ‘LIBELF_BYTEORDER’ undeclared here (not in a function); did you mean
> ‘LIBELF_ERROR’?
>     35 | .libelf_byteorder   = LIBELF_BYTEORDER,
>    |   ^~~~
>    |   LIBELF_ERROR
> ../rtemstoolkit/elftoolchain/libelf/elf.c:36:35: error: ‘LIBELF_CLASS’
> undeclared here (not in a function)
>     36 | .libelf_class   = LIBELF_CLASS,
>    |   ^~~~
> 
> Waf: Leaving directory
> 
> `/home/rtems-tester/rtems-cron-6/rtems-source-builder/rtems/build/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86-1/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86/build'
> 
> 
> Does anyone know a fix for this?
> 
> 
> rtemstoolkit/elftoolchain/libelf/_libelf_config.h has ifdef's for various host
> operating
> systems and does not have one for Cygwin or MSYS. But it can't know the 
> native ELF
> details because the native format is not ELF:
> 
> $ file /bin/cat.exe
> /bin/cat.exe: PE32+ executable (console) x86-64, for MS Windows
> 
> I'm not sure what to do about this. My first thought would be to add 
> Cygwin to the list like FreeBSD and see if that defines what you need.
> or find the x86_64 Linux settings and use those for Cygwin to let it
> compile.
> 
> I'm not sure why we need to know the native ELF details.

The tools use to build on Ming64...

https://git.rtems.org/rtems-tools/tree/rtemstoolkit/elftoolchain/libelf/_libelf_config.h?h=5#n191

This patch broke the support...

https://git.rtems.org/rtems-tools/commit/rtemstoolkit/elftoolchain/libelf/_libelf_config.h?id=ff68bccc25386c84d893187af36e4bac88db4b10

Consider a patch to add the piece removed at the end of this patch as approved.

Chris
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: Cygwin tools failure

[Joel Sherrill]

On Tue, Oct 11, 2022 at 8:36 PM Ryan Long  wrote:

> Cygwin has been unable to build the tools for awhile. This has been due
> to an array subscript being a char while building DTC. The maintainers
> didn't see an issue with the code itself, so it's going to go on being
> broken for awhile I guess.
>
> Joel recommended commented out the building of DTC through the RSB.
> After doing so, I'm running into the following error. I was building
> AArch64's tools. Disabling DTC and building with my Debian WSL instance
> built it just fine.
>
>
> [ 48/258] Compiling rtemstoolkit/elftoolchain/libelf/libelf_align.c
> ../rtemstoolkit/elftoolchain/libelf/elf.c:34:35: error: ‘LIBELF_ARCH’
> undeclared here (not in a function); did you mean ‘LIBELF_ERROR’?
> 34 | .libelf_arch= LIBELF_ARCH,
>|   ^~~
>|   LIBELF_ERROR
> ../rtemstoolkit/elftoolchain/libelf/elf.c:35:35: error:
> ‘LIBELF_BYTEORDER’ undeclared here (not in a function); did you mean
> ‘LIBELF_ERROR’?
> 35 | .libelf_byteorder   = LIBELF_BYTEORDER,
>|   ^~~~
>|   LIBELF_ERROR
> ../rtemstoolkit/elftoolchain/libelf/elf.c:36:35: error: ‘LIBELF_CLASS’
> undeclared here (not in a function)
> 36 | .libelf_class   = LIBELF_CLASS,
>|   ^~~~
>
> Waf: Leaving directory
>
> `/home/rtems-tester/rtems-cron-6/rtems-source-builder/rtems/build/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86-1/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86/build'
>
>
> Does anyone know a fix for this?
>

rtemstoolkit/elftoolchain/libelf/_libelf_config.h has ifdef's for various
host operating
systems and does not have one for Cygwin or MSYS. But it can't know the
native ELF
details because the native format is not ELF:

$ file /bin/cat.exe
/bin/cat.exe: PE32+ executable (console) x86-64, for MS Windows

I'm not sure what to do about this. My first thought would be to add
Cygwin to the list like FreeBSD and see if that defines what you need.
or find the x86_64 Linux settings and use those for Cygwin to let it
compile.

I'm not sure why we need to know the native ELF details.

--joel


> Thanks,
>
> Ryan
>
> ___
> devel mailing list
> devel@rtems.org
> http://lists.rtems.org/mailman/listinfo/devel
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Cygwin tools failure

[Ryan Long]

Cygwin has been unable to build the tools for awhile. This has been due 
to an array subscript being a char while building DTC. The maintainers 
didn't see an issue with the code itself, so it's going to go on being 
broken for awhile I guess.


Joel recommended commented out the building of DTC through the RSB. 
After doing so, I'm running into the following error. I was building 
AArch64's tools. Disabling DTC and building with my Debian WSL instance 
built it just fine.



[ 48/258] Compiling rtemstoolkit/elftoolchain/libelf/libelf_align.c
../rtemstoolkit/elftoolchain/libelf/elf.c:34:35: error: ‘LIBELF_ARCH’ 
undeclared here (not in a function); did you mean ‘LIBELF_ERROR’?

   34 | .libelf_arch    = LIBELF_ARCH,
  |   ^~~
  |   LIBELF_ERROR
../rtemstoolkit/elftoolchain/libelf/elf.c:35:35: error: 
‘LIBELF_BYTEORDER’ undeclared here (not in a function); did you mean 
‘LIBELF_ERROR’?

   35 | .libelf_byteorder   = LIBELF_BYTEORDER,
  |   ^~~~
  |   LIBELF_ERROR
../rtemstoolkit/elftoolchain/libelf/elf.c:36:35: error: ‘LIBELF_CLASS’ 
undeclared here (not in a function)

   36 | .libelf_class   = LIBELF_CLASS,
  |   ^~~~

Waf: Leaving directory 
`/home/rtems-tester/rtems-cron-6/rtems-source-builder/rtems/build/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86-1/rtems-tools-d0a65c72d1a170637258eb19f7d3e433be7c3c86/build'



Does anyone know a fix for this?

Thanks,

Ryan

___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel