On 6/27/24 14:31, Richard Purdie wrote:
On Thu, 2024-06-27 at 12:34 +0000, Ross Burton wrote:
On 27 Jun 2024, at 13:02, Richard Purdie
<[email protected]> wrote:
On Wed, 2024-06-26 at 10:41 +0100, Peter Hoyes via
lists.openembedded.org wrote:
It does not always make sense to collect artifacts and data from
the
target on failure, e.g. if testing firmware or if the target is
not
running an SSH server.
Introduce the variable TESTIMAGE_RUN_FAILURE_POST_ACTIONS, which
defaults to "1". If this variable is not true, skip the failed
test post
actions.
Signed-off-by: Peter Hoyes <[email protected]>
---
meta/classes-recipe/testimage.bbclass | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)
Is there a reason you can't just set TESTIMAGE_FAILED_QA ARTIFACTS
= "”
?
Answering for Peter because I asked the same question: the next step
after list_and_fetch_failed_tests_artifacts() is
get_target_disk_usage() which does a target.run() so causes a SSH
call.
Rather than call this RUN_FAILURE_POST_ACTIONS could we have something
like TESTIMAGE_SSH_AVAILABLE and then test against that in the
appropriate places?
Cheers,
Richard
In our test suite at least, SSH is sometimes available and we use SSH
where it makes sense. We just can't guarantee that SSH is available in
all failure scenarios (e.g. if the firmware failed to boot), and many of
our test cases run pre-Linux shell.
Peter
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#201205):
https://lists.openembedded.org/g/openembedded-core/message/201205
Mute This Topic: https://lists.openembedded.org/mt/106886598/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
