On 04/04/18 03:42 PM, Paul Smith wrote:
On Wed, 2018-04-04 at 13:03 -0400, Dennis Clarke wrote:
After all the vibrant discussion I was at least expecting a reply that
says "okay .. so that works" or perhaps a "ver 4.2.2 patches?" or
Well, we thought it would work and it did work, and those fixes are in
the codebase... so win!
It does look like we need to make a new release soon.
A minor release anyways.
What bothers me is that these patches are only needed on a i686 system
I find that difficult to understand. Are you SURE you're using the
exact same GNU libc on all the different systems?
Ignore me .. that is wrong. I did have make 4.2.1 on multiple systems
but I went back and did an update in a few places and some tests and
the make I had previously built was with :
nix$ readelf -delVt `whicjh gmake`
Version needs section '.gnu.version_r' contains 2 entries:
Addr: 0x0000000010004a78 Offset: 0x004a78 Link: 6 (.dynstr)
000000: Version: 1 File: libdl.so.2 Cnt: 1
0x0010: Name: GLIBC_2.3 Flags: none Version: 5
0x0020: Version: 1 File: libc.so.6 Cnt: 3
0x0030: Name: GLIBC_2.17 Flags: none Version: 4
0x0040: Name: GLIBC_2.4 Flags: none Version: 3
0x0050: Name: GLIBC_2.3 Flags: none Version: 2
ah ha ... well now I have this :
nix$ ldd --version
ldd (Debian GLIBC 2.27-3) 2.27
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.
Yes the problem exists.
As I pointed out in a previous post, this has no relationship to the
underyling architecture, compiler, etc. As far as I understanding it
this issue comes solely from the version of GNU libc you're using.
Seems to be exactly the case too.
Note those patches are not the same as what you have in your links to
git diff reports. I had to make changes to get them to work against the
Bug-make mailing list