On Wed, Apr 11, 2018 at 3:09 AM, Alexander Kanavin
<alexander.kana...@linux.intel.com> wrote:
> On 04/11/2018 12:20 AM, Andre McCurdy wrote:
>>> -#| engines/afalg/e_afalg.c: In function 'eventfd':
>>> -#| engines/afalg/e_afalg.c:110:20: error: '__NR_eventfd' undeclared
>>> (first use in this function)
>>> -#|      return syscall(__NR_eventfd, n);
>>> -#|                     ^~~~~~~~~~~~
>>> -EXTRA_OECONF_aarch64 += "no-afalgeng"
>>> +# This prevents openssl from using getrandom() which is not available on
>>> older glibc versions
>>> +# (native versions can be built with newer glibc, but then relocated
>>> onto a system with older glibc)
>>> +EXTRA_OECONF_class-native += "--with-rand-seed=devrandom"
>>> +EXTRA_OECONF_class-nativesdk += "--with-rand-seed=devrandom"
>> Better to avoid += with an over-ride since it doesn't do what most new
>> users etc expect. Better to use _append instead (or just the over-ride
>> on it's own, if over-riding the original value is what you intended to
>> do).
> Thanks, I will fix this. The whole += vs. _append is a design shortcoming in
> bitbake I'd say, as it's a regular source of confusion, and can't be
> automatically QA'd.

Yes, fully agree. We would save new users a lot of time and confusion
if the parser just failed with an error on any attempt to use += with
an over-ride. I've seen it cause bugs many times.
Openembedded-core mailing list

Reply via email to