Bugzilla – Full Text Bug Listing
|Summary:||raise glibc requirements|
|Product:||ThinLinc||Reporter:||Pierre Ossman <email@example.com>|
|Component:||Build system||Assignee:||Pierre Ossman <firstname.lastname@example.org>|
|Status:||CLOSED FIXED||QA Contact:||Bugzilla mail exporter <email@example.com>|
|Version:||pre-1.0||Keywords:||pcok, relnotes, samuel_tester|
|Bug Depends on:|
|Bug Blocks:||5746, 7196|
At some point we should consider raising our requirements on glibc so that we can make use of newer features. For the server this is probably not a big issue as our requirements on other things forces a new glibc anyway. The big problem is the terminals which might be based on ancient distributions.
I did an inventory and raising the requirement to 2.11.3 will still handle all interesting terminals. This unfortunately doesn't get us the memcpy() performance improvement (2.14), but it does mean we can have the same glibc for all our build environments. It will probably also solve various minor annoyances. For reference, it is Wyse SUSE Linux 11 that has the oldest glibc. However RHEL 6 has 2.12 so we couldn't go much further even without Wyse. Third oldest is RHEL 7 with 2.17.
The old Wyse SUSE Linux 11 is on its way out anyway as we'll start looking at it's replacement soon (bug 6199). Dell still claims these devices aren't EOL, but they haven't been sold in years. Let's not let these devices hold us back and we'll use RHEL 6 as a base for everything.
The upstream SLES 11 is also affected by this. Support for it ends in March 2019. That is not long after our next release is out, so it's not really a problem dropping support for it.
Should be all done. I cannot find any more references to wyse or glibc that need changing. I also tested the autotest by putting some Fedora binaries in /opt/thinlinc/bin.