Re: [OE-core] [PATCH v7 8/9] oe-selftest/cases/wic.py: oe-selftest -r wic.Wic -> PASS

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

On Sun, 2026-03-01 at 19:38 -0500, Trevor Woerner via lists.openembedded.org wrote:
> Changes for the wic.Wic oe-selftests to pass now that wic is a
> stand-alone app and no longer part of oe-core.
> 
> Signed-off-by: Trevor Woerner <twoerner@gmail.com>
> ---
> changes in v7:
> - none
> 
> changes in v6:
> - none
> 
> changes in v5:
> - rebase with master
> - split back out into multiple, smaller patches
> - now that the sector-size re-implementation is ahead of this patch,
>   several wic subcommands (ls, cp, rm) no longer need a --vars file
> 
> v4: skipped to align/combine with other patch set
> 
> changes in v3:
> - squashed into 1 large, atomic commit
> 
> changes in v2:
> - none
> ---
>  meta/lib/oeqa/selftest/cases/wic.py | 136 ++++++++++++++++------------
>  1 file changed, 76 insertions(+), 60 deletions(-)

I mentioned this on irc but for completeness, I should respond here too.

The commit message says "what" the patch does but nothing about why,
there is nothing explaining the changes.

For a major migration like this, the explanation of why is important,
it may well help other people convert their code for example.

In this case, the issue (at least as I understand it) is that the
external wic tool cannot call bitbake to extract bitbake variable
values any longer so we need to add the environment file to the wic
calls to make sure it has that information. This should be mentioned in
the commit message but it isn't. I shouldn't have to go and ask people
and spend an age reading the code.

In some ways, this patch would be better as a standalone patch before
the split out, that way we could test before and after the split and
show no regressions and it would make the intent of the change clearer.

To be clear, I don't object to removing the support for calling bitbake
to get variable values. What I don't like is the way this is being done
with no visibility to the change.

In digging into this, I also noticed other issues:

* the wic repo has both master and main branches. This is extremely
confusing and it shouldn't have both, particularly when the content is
totally different

* there is no explanation of the changes in the wic repo to make it
external. The removal of support for executing bitbake for example is
never mentioned there either. Where were those patches reviewed? What
will the review process be in future?

* there is a workaround for e2fsck return codes in the wic repo. Where
was this posted for review? Is that a change we'd have accepted in OE-
Core? I suspect not as I have serious concerns about a workaround like
that. Unfortunately that raises questions about the ongoing maintenance
of the repo too.

Cheers,

Richard