On 2015-08-24 21:39, Jacco Ligthart wrote:
Jacco, have you had any success in getting the EL6 version to build?
Anyone
else?
Nope, the last version I tried was 31.4 (standard EL6). my notes say
that the error was:
Architecture-arm.cpp:101: error: 'HWCAP_VFPv4' was not declared in
this
scope
I remember that, there were assumptions made about ARM and FPU.
I've had a patch for that in ages, it should be on the ftp server.
OK I'll have a look at that.
It's quite an old patch, but have a look here:
http://ftp.redsleeve.org/pub/patches/firefox-24.2.0-1.el6_5.src.rpm.patch
It adds a distinction between ARMs with and without SIMD FPU.
This seems to be a very long standing bug/deficiency in upstream
FF code, but given to what extent they cannot be arsed with
supporting anything but their latest agilely hacked together
code, I am not sure I can muster the energy to try to push it
upstream.
FYI, I'm currently busy rebasing EL6 to upstream version 6.7
I'm getting close to the point (I hope) where I can start the bulk
build. I do however see that there are again more patches needed to
get
it al to work.
Yes, I got a few dependencies updated to build firefox, but FF
is being really problematic. >:(
I meant more that python now needs patched to build, and one or two
other.
Ah, now that you mention it I do remember this python issue. :(
Notable exception: the kernel. It now creates without any patches a
kernel-headers package :)
Wait - what? Are you saying the latest EL6 upstream kernel builds
on ARM? What config/SoC?
only headers, so no "real" kernel, so there is no config and no mention
of a SOC. But I was surprised to see that there was mention of %{arm}
in
the spec and that it produced something.
Ah, sorry, I misunderstood. Headers only makes more sense.
This is a bit of a problem since the last version that I managed to
get to build on RSEL6 has some serious security issues.
You mean the kernel? The last EL6 I managed to get to build from
the upstream src.rpm with a custom config and a fair few patches
was 2.6.32-200.x series.
I don't get your remark here. I think you're talking to yourself :)
Sorry, I hadn't spotted the distinction you were making between
kernel-headers and the full kernel package.
Gordan
_______________________________________________
users mailing list
[email protected]
http://lists.redsleeve.org/mailman/listinfo/users
