hi there,
as discussed recently, we are starting to put together the thud builds for Linaro OE RPB.
* The oe-rpb-manifest now has the 'thud' branch, see [1]. Some layers are still missing thud branch themselves, in which case we use master. * we created thud branches for all Linaro layers (meta-linaro, meta-backports, meta-rpb, meta-96boards, meta-qcom) * A new build was created on ci.linaro, see [2] , and is now enabled (which means it will trigger on any change on one of the thud branches we track)
The first 2 builds failed during the weekend, but it was infrastructure issues, however we just got a new failure that is a valid failure.. see [3]. Looks like LTP fails to build.
Daniel: is that something you can check and fix?
cheers nico
[1] https://github.com/96boards/oe-rpb-manifest/tree/thud [2] https://review.linaro.org/#/c/ci/job/configs/+/29110/ [3] https://ci.linaro.org/job/96boards-reference-platform-openembedded-thud/
Hello!
On Mon, 19 Nov 2018 at 17:48, Nicolas Dechesne nicolas.dechesne@linaro.org wrote: [...]
Looks like LTP master is currently failing, an unfortunate coincidence. However, I find it odd that Bitbake is preferring that version (ltp_git.bb) versus the one provided by OE-core (ltp_20180926.bb).
Daniel: is that something you can check and fix?
Yes. I already build-tested Thud + LTP 20180926 and have a patch ready for meta-backports, which might become a special case for a few days, noting that a few things are missing from the LTP recipe: * RDEPENDS that we have in our meta-backports which are not in OE-core, and * an additional patch that needs to be dropped from OE-core.
Once those two things happen in OE-core, we'll resync to all backports (Thud, Sumo Rocko, and Morty). In the meantime, Thud will differ slightly from the other three branches, which already have a backport.
Greetings!
Daniel Díaz daniel.diaz@linaro.org