On Thu, 2026-07-30 at 09:59 -0500, Ryan Eatmon via lists.openembedded.org wrote: > > > On 7/20/2026 12:11 PM, Freihofer, Adrian wrote: > > > > Hi Yoann > > > > Just a confirmation from my side: Taking the 4 patches is the best > > we > > can do. Considering them as a series of bug fixes is right. Getting > > the > > test back-ported is of course nice to have. > > > > Regards > > Adrian > > > Backporting the test will be challenging as I built the test on top > of > the 7 patch series from you that overhauled a lot of the fitImage > testing. > > https://lore.kernel.org/openembedded-core/[email protected]/ > https://lore.kernel.org/openembedded-core/[email protected]/ > https://lore.kernel.org/openembedded-core/[email protected]/ > https://lore.kernel.org/openembedded-core/[email protected]/ > https://lore.kernel.org/openembedded-core/[email protected]/ > https://lore.kernel.org/openembedded-core/[email protected]/ > > That means that my backport would need to be making a bunch of > changes > that these patches may have done. I can do that, if everyone wants > me > to. I know specifically that this patch is a big one that mine is > making use of: > > https://git.openembedded.org/openembedded-core/commit/meta/lib/oeqa/selftest/cases/fitimage.py?id=4b08cf6048b6575c39aa9b3bea282cdeae228e8a >
I agree that this is not very easy. It would require a pretty wrynose specific patch. This is also sort of developing on stable. On the existing tests on wrynose there is coverage for FIT_CONF_DEFAULT_DTB for DTB names without a path. Since the beaglebone is the reference, there is only FIT_CONF_DEFAULT_DTB = "am335x-bonegreen.dtb" https://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/selftest/cases/fitimage.py?h=wrynose#n983 Adding test coverage on master is much simpler because there we can also find a reference machine with a default DTB like "arm/foundation- v8.dtb" https://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/selftest/cases/fitimage.py#n120 @Yoann: The missing detail is test coverage for default DTBs with a directory in the name. All the rest seems to be covered by the existing tests. Alternative would be to go with only three patches. From a users perspective that would mean a breaking change with FIT_CONF_DEFAULT_DTB when updating from wrynose to a later version. With the four patches it will be future proof from now onward. So my opinion is: Let's go with four patches and 99% coverage for wrynose. That's good enough. Regards, Adrian > > > > > > > > >
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#242381): https://lists.openembedded.org/g/openembedded-core/message/242381 Mute This Topic: https://lists.openembedded.org/mt/120329199/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
