Re: [Xen-devel] [OSSTEST PATCH] ts-depriv-audit-qemu: Leaked substeps are `blocked', and then script is fail

[Anthony PERARD <anthony.perard@xxxxxxxxxx>]

On Fri, Jul 06, 2018 at 12:13:54PM +0100, Ian Jackson wrote:
> If thus script crash in some non-substep-specific operation, the
> substeps may have been created (and thus be in the db as `running').
> Leaving them like this is not allowed: it is treated by sg-run-job,
> and by the archaeologists as an internal error.
> 
> If this happens, what we want is for the script as a whole to be
> `fail', and for the failing tests to be `blocked'.  The former will
> spot if this crash is a regression.  The latter will avoid treating
> the substeps themselves as regressions.
> 
> Reported-by: Anthony PERARD <anthony.perard@xxxxxxxxxx>
> Signed-off-by: Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>
> ---
>  ts-depriv-audit-qemu | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/ts-depriv-audit-qemu b/ts-depriv-audit-qemu
> index 81bd5c0..3c218d9 100755
> --- a/ts-depriv-audit-qemu
> +++ b/ts-depriv-audit-qemu
> @@ -118,10 +118,21 @@ sub audit_fish () {
>      foreach my $cl (sort keys %classes) {
>       substep_finish("/$cl", $classes{$cl} || 'pass');
>       # (if there are no fds of this class, maybe they were all closed)
> +     $classes{$cl} = 'reported';

The case here is different than ...

>      }
>      logm("audit complete...");
>  }
>  
> +END {
> +    foreach my $cl (sort keys %classes) {
> +     my $st = $classes{$cl};
> +     next if $st eq 'REPORTED';

.. here. Is there something I'm missing or should it be the same?


-- 
Anthony PERARD

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel