On Wed, 2023-06-07 at 14:29 +0200, Alexander Kanavin wrote: > On Wed, 7 Jun 2023 at 14:23, Alexis Lothoré <[email protected]> > wrote: > > > Then core-image-ptest.bb should append "${libdir}/{MCNAME}/ptest" to > > > that. As that same image recipe installs only a single ptest via > > > IMAGE_INSTALL:append, and the artifacts are retrieved only if that > > > ptest would fail, this would achieve the same outcome. Other images > > > (which can potentially include many ptests) can then set what > > > artifacts to retrieve themselves as they please. > > > > I am not familiar with multiconfig, so I'll have to document myself before > > trying what you suggest, especially to make sure it will not make the > > retriever > > too dependant on a proper multiconfig definition. > > core-image-ptest.bb is already set for this, so it just adds another > usage of MCNAME variable into it to get the name of the ptest from the > mcextend variant and substitute that into the path that gets added to > TESTIMAGE_FAILED_QA_ARTIFACTS. The retriever won't know anything about > it as it happens at the recipe parsing step.
It does seem like a neat way to handle things to me FWIW. Cheers, Richard
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#182475): https://lists.openembedded.org/g/openembedded-core/message/182475 Mute This Topic: https://lists.openembedded.org/mt/99380468/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
