Sebastian Ramacher wrote:
Thank you for the explanation. What about the attached patch? [1]
I've looked through the code and the only place where --enable-runtime-cpudetect
makes a difference is on powerpc (libavutil/ppc/cpu.c). Special handling for
Raspbian should not make a difference. Reinhard, is my reading of the code
correct?
I honestly don't recall why exactly I enabled that option. It's possible
I determined it was needed at the time I first prepared the patch (which
was some time ago, I generally forward port changes rather than starting
from scratch each time), it's also possible that I enabled it and it
didn't stop the package coming out with armv7 code in it but I never
disabled it again. Lets try without it and see what happens.
[1] Is this the right vendor in dpkg-vendor's terms?
It's the right vendor but in general it's better to use --derives-from
rather than --is in situations like this on the principle that a newly
formed derivative should as much as possible have the same behaviour as
the distro it derives from.
I've applied your patch, tweaked it to use --derives-from rather than
--is implemented this change, written a changelog entry (which may be
more verbose than actually needed, feel free to cut it down if you want
when bringing the change into Debian) and am now running a build.
Assuming it builds successfully and passes our contamination checker I
will be uploading it to raspbian jessie.
I suspect that the release team will nack this change for Debian jessie
so uploading this change to Debian will probablly have to wait until
after the jessie release (or go to experimental).
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
