Re: [PATCH v5 11/17] tests/docker: strip remaining caching logic from docker.py

["Alex Bennée" <alex.bennee@linaro.org>]

Pierrick Bouvier <pierrick.bouvier@oss.qualcomm.com> writes:

> On 9/17/2026 9:55 AM, Alex Bennée wrote:
>> The caching logic was a remnant of an older time when we had multiple
>> shared layers building that went into the final image. It was arguable
>> that it worked back then but it is definitely not needed now so into
>> the bin it goes.
>> 
>> Tested-by: Richard Henderson <richard.henderson@linaro.org>
>> Reviewed-by: Pierrick Bouvier <pierrick.bouvier@oss.qualcomm.com>
>> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
>> ---
>>  tests/docker/docker.py | 148 ++++++++++++++---------------------------
>>  1 file changed, 49 insertions(+), 99 deletions(-)
>>
>
> After testing this series, it's not a good idea to include this patch.
>
> It forces a systematic rebuild, even when image is already. This has
> three consequences:
> - verbosity: we see build log systematically

You asked me to drop the verbose/quiet switch patch which addresses this.

> - slower: running a noop rebuild takes time

how much time? it flies past when I run a rebuild.

> - network dependency: docker.py --registry triggers a pull, which fails
> if there is no network. It breaks offline make check.
> Also, it slows down things too.

we should be able to gate the registry on network access.

>
> So, let's drop this patch and keep the caching mechanism for now, it's
> still useful/needed.
>
> Regards,
> Pierrick

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro