Client: Keybase ui segfault after glibc upgrade on Antergos glibc 2.28-1

[mikaelkall]

Keybase ui segfault after glibc upgrade on Antergos glibc 2.28-1

my log id: 8fcb994c0887962aaba6d41c

[AWildBeard]

I can confirm I'm getting the same issue under Arch Linux.
keybase-bin 2.6.0_20180815194228+604cfa0d29-1

my logs: 2505b18b544b1edbd6b64d1c

UI logs were not included because, well, the UI segfaults on launch.
But here's a GDB backtrace if that helps :D

Program received signal SIGSEGV, Segmentation fault.
0x0000000000b87650 in ?? ()
(gdb) bt
#0  0x0000000000b87650 in  ()
#1  0x00007ffff7d833e2 in node::http2::Http2Session::Callbacks::Callbacks(bool) () at /opt/keybase/libnode.so
#2  0x00007ffff7d834a5 in  () at /opt/keybase/libnode.so
#3  0x00007ffff7fe36da in call_init.part () at /lib64/ld-linux-x86-64.so.2
#4  0x00007ffff7fe37da in _dl_init () at /lib64/ld-linux-x86-64.so.2
#5  0x00007ffff7fd503a in _dl_start_user () at /lib64/ld-linux-x86-64.so.2
#6  0x0000000000000001 in  ()
#7  0x00007fffffffe953 in  ()
#8  0x0000000000000000 in  ()

[jerryaldrichiii]

Also seeing the same with keybase-bin 2.6.0_20180816164108+2154e265f3-1

[AWildBeard]

I updated again, got a kernel update and a glibc update to 2.28-4.

It is now working fine, but I see on the AUR that there is indeed a newer keybase I need to update to. I'm in the middle of a project right now but I will see if I can upgrade later and report back. Think this might be a glibc problem that just got fixed today on the glibc end? I.E. glibc 2.28-4?

[AWildBeard]

Upgraded keybase to:
keybase-bin 2.6.0_20180817210538+0b2dcf191f-1

And everything seems to be working just fine.

Like a phantom this segfault has gone away as fast as it came for me.

[jerryaldrichiii]

I can also confirm that upgrading to glibc-2.28-4 and keybase-bin_2.6.0_20180817210538+0b2dcf191f-1 fixed the issue.

What a strange couple of days @AWildBeard.

[mikaelkall]

Resolved

[eli-schwartz]

This is really just #13180 and therefore https://github.com/electron/electron/issues/13972 (an error in electron itself generating invalid code).

It's worked around by the glibc update, and fixed properly by the electron 2.0.8 update.