In case the revert
https://sourceware.org/ml/libc-alpha/2016-06/msg00339.html
proceeds upstream, I am going to turn recvmsg@GLIBC_2.24 and the related symbols into a compat symbol, and move the default back to the previous version. This should keep existing binaries working, while new binaries will pick up the previous symbol version.
After a while, I'll use symboldb https://pagure.io/symboldb to check if there are still any RPMs
So in short, we have a plan to implement the revert even without a mass rebuild. You don't have to oppose the revert for Fedora's sake. :)
Florian
On 06/10/2016 10:00 AM, Florian Weimer wrote:
In case the revert
https://sourceware.org/ml/libc-alpha/2016-06/msg00339.html
proceeds upstream, I am going to turn recvmsg@GLIBC_2.24 and the related symbols into a compat symbol, and move the default back to the previous version. This should keep existing binaries working, while new binaries will pick up the previous symbol version.
After a while, I'll use symboldb https://pagure.io/symboldb to check if there are still any RPMs
remaining which reference the new GLIBC_2.24 symbols, I'll ask provenpackagers to do targeted rebuilds.
So in short, we have a plan to implement the revert even without a mass rebuild. You don't have to oppose the revert for Fedora's sake. :)
Florian
On 06/10/2016 04:23 AM, Florian Weimer wrote:
On 06/10/2016 10:00 AM, Florian Weimer wrote:
In case the revert
https://sourceware.org/ml/libc-alpha/2016-06/msg00339.html
proceeds upstream, I am going to turn recvmsg@GLIBC_2.24 and the related symbols into a compat symbol, and move the default back to the previous version. This should keep existing binaries working, while new binaries will pick up the previous symbol version.
After a while, I'll use symboldb https://pagure.io/symboldb to check if there are still any RPMs
remaining which reference the new GLIBC_2.24 symbols, I'll ask provenpackagers to do targeted rebuilds.
So in short, we have a plan to implement the revert even without a mass rebuild. You don't have to oppose the revert for Fedora's sake. :)
The revert has gone in, so we need to fix this on our end to keep binaries running until they get rebuilt.
I don't oppose the revert. Given the problems encountered it's the right thing to do. I still think that staying close to the standards is important.