Re: [PATCH 2/2] buildman: Add --allow-missing-binaries flag to build with BINMAN_ALLOW_MISSING=1

[Tom Rini <trini@konsulko.com>]

[-- Attachment #1: Type: text/plain, Size: 2283 bytes --]

On Mon, Oct 10, 2022 at 05:48:55PM -0600, Simon Glass wrote:
> Hi Tom,
> 
> On Mon, 10 Oct 2022 at 09:18, Tom Rini <trini@konsulko.com> wrote:
> >
> > Add a new flag to buildman so that we will in turn pass
> > BINMAN_ALLOW_MISSING=1 to 'make'. Make use of this flag in CI.
> >
> > Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
> > Cc: Simon Glass <sjg@chromium.org>
> > Signed-off-by: Tom Rini <trini@konsulko.com>
> > ---
> >  .azure-pipelines.yml            | 2 +-
> >  .gitlab-ci.yml                  | 6 +++---
> >  tools/buildman/builder.py       | 5 ++++-
> >  tools/buildman/builderthread.py | 2 ++
> >  tools/buildman/cmdline.py       | 3 +++
> >  tools/buildman/control.py       | 3 ++-
> >  6 files changed, 15 insertions(+), 6 deletions(-)
> >
> > diff --git a/.azure-pipelines.yml b/.azure-pipelines.yml
> > index f200b40dbb24..c932c2b3c619 100644
> > --- a/.azure-pipelines.yml
> > +++ b/.azure-pipelines.yml
> > @@ -553,7 +553,7 @@ stages:
> >            cat << "EOF" >> build.sh
> >            if [[ "${BUILDMAN}" != "" ]]; then
> >                ret=0;
> > -              tools/buildman/buildman -o /tmp -P -E -W ${BUILDMAN} ${OVERRIDE} || ret=$?;
> > +              tools/buildman/buildman -o /tmp -P -E -W --allow-missing-binaries ${BUILDMAN} ${OVERRIDE} || ret=$?;
> 
> This is fine for CI.

I think one of the issues here is we need to agree on the common use
cases. I don't think most people use buildman to build, they use make.
Of the people that do use buildman, how many aren't already using a
wrapper script? I know I always do. Can we also not just deal with
setting this flag in ~/.buildman ? Would it Just Work as:
[builder]
allow-missing-binaries = True

Or is more code needed?

> But having to provide a flag to do build testing seems like the tail
> is wagging the dog. Boards should be discouraged to use blobs and we
> don't want to make it hard for everyone else (who doesn't have the
> blobs) to test whether their patch breaks a build.

I'm not sure this ends up being a common case. If someone is build
testing for something they can't run test, the error is going to be
after whatever they compile-tested was compiled/linked, and if they're
testing everything it's via CI.

-- 
Tom

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]