Re: [docs] "Required Variables" aren't really "required"

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Fri, 2026-06-05 at 12:32 -0400, Robert P. J. Day wrote:
>   yup, and i was a bit surprised at the result (although i'm sure this
> will come as no surprise to you).
> 
>   with wrynose release, i picked a random oe-core recipe (util-linux),
> and edited the associated ptest.patch file to delete the line:
> 
>   Upstream-Status: Pending
> 
> i was surprised that if i simply wanted to generate the environment or
> parse the recipe:
> 
>   $ bitbake -e util-linux
>   $ bitbake -p util-linux
> 
> that worked fine, so i guess neither of those operations checks the
> patches.
> 
>   predictably, this command:
> 
>   $ bitbake -c do_patch util-linux
> 
> generated:
> 
>   ERROR: util-linux-2.41.3-r0 do_patch: QA Issue: Missing Upstream-Status in patch
> 
> so that works just fine.
> 
>   in short, the necessity for upstream status applies only to oe-core
> patches. should the docs mention that distinction?

I went and checked and "patch-status" is being added to ERROR_QA
unconditionally so it should apply to other layers too, unless you've
something changing that locally (a layer or local config).

I do feel this discussion is off in the weeds but I really don't like
the docs being altered based on incorrect information.

Cheers,

Richard