Hi,
As you probably know, upstream RHEL 7.3 was released this week. Parallel to the x86_64 rebuild, I wanted to kick the armhfp rebuild. As always, we start by gcc, then glibc, and then the rest. While gcc / binutils built fine (see http://armv7.dev.centos.org/repodir/c71611-pass-1/ for the artifacts and http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/ for the build logs) , glibc actually fails :
http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/15679-glibc-2.17-157.el7/...
That one has to be fixed before launching the massive rebuild. I'm myself in "offline" mode those days, and I'll be in Paris next week for the CentOS Interlock, so if you want to help, have a look at those failing pkgs and possible root cause explanations (and patches ?) and report your findings on this list ?
Cheers,
Hi,
I saw the same thing on ARMv5. Here my patch to make it build.
Jacco
On 06-11-16 09:10, Fabian Arrotin wrote:
Hi,
As you probably know, upstream RHEL 7.3 was released this week. Parallel to the x86_64 rebuild, I wanted to kick the armhfp rebuild. As always, we start by gcc, then glibc, and then the rest. While gcc / binutils built fine (see http://armv7.dev.centos.org/repodir/c71611-pass-1/ for the artifacts and http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/ for the build logs) , glibc actually fails :
http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/15679-glibc-2.17-157.el7/...
That one has to be fixed before launching the massive rebuild. I'm myself in "offline" mode those days, and I'll be in Paris next week for the CentOS Interlock, so if you want to help, have a look at those failing pkgs and possible root cause explanations (and patches ?) and report your findings on this list ?
Cheers,
Arm-dev mailing list Arm-dev@centos.org https://lists.centos.org/mailman/listinfo/arm-dev
On 06/11/16 11:05, Jacco Ligthart wrote:
Hi,
I saw the same thing on ARMv5. Here my patch to make it build.
Hi Jacco,
Thanks a lot .. still almost in "offline" mode now, but I added your patch and submitted a test build : http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/15680-glibc-2.17-157.el7....
Let's see how that goes but what does your patch currently do ?
On 06-11-16 16:43, Fabian Arrotin wrote:
On 06/11/16 11:05, Jacco Ligthart wrote:
Hi,
I saw the same thing on ARMv5. Here my patch to make it build.
Hi Jacco,
Thanks a lot .. still almost in "offline" mode now, but I added your patch and submitted a test build : http://armv7.dev.centos.org/rpmbuild/c71611-pass-1/15680-glibc-2.17-157.el7....
Let's see how that goes but what does your patch currently do ?
Good question. It is more deduction from the circumstances, than that I can tell exactly whit this does.
- The previous version worked, this one does not, it should be one of the new patches that make the FTBFS. - upstream vendor often patches only the x86 part of the code, breaking other archs - I found a patch (glibc-rh1256317-4.patch) with this text: Remove IS_IN_rtld
Replace with IS_IN (rtld) - I guess they did not apply this to all arm parts of the code - the official glibc git has the same line in this place as my patch - it now builds :)
All in all, I'm fairly confident that this does not break things, but I would also like some feedback on this patch.
Jacco
-
Fabian Arrotin -
Jacco Ligthart