On Wed, Aug 23, 2017 at 1:18 PM, Andrei Gherzan <[email protected]> wrote:
> On Wed, Aug 23, 2017 at 6:08 PM, Bruce Ashfield <[email protected]> > wrote: > >> >> >> On Wed, Aug 23, 2017 at 1:01 PM, Andrei Gherzan <[email protected]> >> wrote: >> >>> On Wed, Aug 23, 2017 at 5:54 PM, Bruce Ashfield < >>> [email protected]> wrote: >>> >>>> Already discussed with Khem and Richard in the original >>>> thread. >>>> >>>> I'm sending another patch shortly to back this off to a note/warning. >>>> >>>> >>> That is nice but wouldn't that clutter the build logs for no reason? For >>> people who don't take advantage of scc definitions that is. Could you point >>> me to the initial discussion? Just want to understand and maybe vote to >>> remove it altogether. >>> >> >> It's a yocto bug. We need to flag BSP definitions that aren't valid, >> otherwise the tools >> will end up building things like qemuarm for someone that has a >> misconfigured BSP. >> >> If you don't have a defconfig (as someone won't have if they are using a >> fully defined >> BSP, but with the wrong name) and the wrong entry point, or no entry >> point, is returned >> we have to flag it. >> >> The actual bug is: 11878, and there's a related fallout of that bad >> definition. >> >> So no, it won't be deleted, but I will make it check for a few more >> parameters before >> erroring or logging. >> > > That makes sense. Thanks for clarifying. I'll keep an eye on the ml for > the fix. > It'll arrive shortly, testing it now. It shouldn't put anything in your logs during the build now, I'm just building the rpi3 to confirm and re-testing a case that I want to fail :D Bruce > > -- > Andrei Gherzan > -- "Thou shalt not follow the NULL pointer, for chaos and madness await thee at its end"
-- _______________________________________________ Openembedded-core mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-core
