[OE-core] [PATCH] gcc-7.inc: add new warning "Wnot-cross-compiler"

Khem Raj raj.khem at gmail.com
Thu Jul 20 19:44:53 UTC 2017


On Thu, Jul 20, 2017 at 3:40 PM, Andre McCurdy <armccurdy at gmail.com> wrote:
> On Thu, Jul 20, 2017 at 12:18 PM, Burton, Ross <ross.burton at intel.com> wrote:
>>
>> On 20 July 2017 at 19:51, Khem Raj <raj.khem at gmail.com> wrote:
>>>
>>> it does cause harm, by making the behaviour inconsistent and highly
>>> dependent
>>> upon build host.
>>
>> It *stops* behaviour being dependent on the build host, surely.  The problem
>> is that some recipes use the host compiler when they're not meant to be and
>> this can cause subtle (or not so subtle) breakage.
>
> Is there an example of a recipe which doing that which went unnoticed?
> Doesn't the fact that most recipes do get cross compiled (ie x86
> target on an x86 host is the corner case) should mean that bugs like
> that get flushed out pretty quickly?

in general yes, most of the issues should already be addressed since we
do cross compile a lot of host != target cases. Case where this might be
not catching are cases where the fall back is cross compiler and first choice
is native gcc. I agree though to have qualified data to see how often does
this occur.



More information about the Openembedded-core mailing list