[PATCHv2 rtems_waf] Allow vendor field in toolchain target triplet

[Chris Johns]

On 25/5/2023 6:53 pm, martinerikwerner.aac at gmail.com wrote:
> While poking around some more, it seems like there's more places in
> this file where assumptions of no vendor in the triplet might come into
> play (but did not affect my use of it), if I'm reading it correctly?:
> 
> 227         conf.env.ARCH_BSP = '%s/%s' % (arch.split('-')[0], bsp)
> 
> 232         conf.env.RTEMS_ARCH = arch.split('-')[0]
> 
> 554     return _arch_from_arch_bsp(arch_bsp).split('-')[0]
> 
> 931         ab = arch_bsp.split('-')
> (...)
> 939         flagstr = subprocess.check_output(
> 940             [config, '--bsp',
> 941              '%s/%s' % (ab[0], ab[2]), flags_map[flags]])
> 
> I'm also a bit uncertain, what is the "arch" actually supposed to be in
> general, given that the _arch_from_arch_bsp(arch_bsp) and
> arch(arch_bsp) seem to disagree (former include the os (rtems) field,
> latter excludes it).

I suspect the uncertainly is due to things evolving and me not paying close
enough attention because it did not matter. The vendor field changes that. I
will take a look and see if I can clean it up.

What is the full text for the `arch_bsp` you are working with so I can test it?

> I'm not sure I am aware of all places where this is used as a
> submodule, my only testing has been as part of a custom user
> application...

Oh nice to hear. I will look through the repos we have and look at updating them.

Chris