Re: [PATCH] scripts/ci_build.sh: Don't fail on bullseye build errors

[Henning Schild <henning.schild@siemens.com>]

On Thu, 30 Jul 2020 18:08:55 +0200
Baurzhan Ismagulov <ibr@radix50.net> wrote:

> On Mon, Jun 22, 2020 at 12:42:51PM +0200, Henning Schild wrote:
> > We need a real fix for this. And if we go for a temporary skip we
> > need a proper testing framework ...
> > 
> > With this hack all we do is potentially introduce more problems for
> > bullseye which will hit us when we have the fix, which we need
> > anyways.  
> 
> I agree that bullseye must be fixed. The problem is upstream, so that
> will take time.

Did you report it upstream? Where is the bug? With you hiding the issue
it is you who needs to do the upstream work.

Henning

> I also agree that we need a proper testing framework. It is actually
> in place. As desired, every scenario has its own test case (a, b, c).
> However, ci_build.sh tests many things in at once (a + b + c). Due to
> that, the test cases don't result in the same coverage. Just adding
> the two would double the time needed (a, b, c, a + b + c). So, I'd
> like to find a good compromise.
> 
> I agree that KFAIL may potentially hide further problems with
> bullseye. However, as you mention in the other thread, it also blocks
> CI -- that is, any work. That is why I think that unblocking CI has
> priority over fixing bullseye.
> 
> Applied to next.
> 
> With kind regards,
> Baurzhan.
>