neovim segfaults

DoneSubmitted by Julien Lepiller.
Details
5 participants
  • Gábor Boskovits
  • Bradley Haggerty
  • Efraim Flashner
  • Julien Lepiller
  • Ricardo Wurmus
Owner
unassigned
Severity
normal
J
J
Julien Lepiller wrote on 22 Feb 2019 14:06
(address . bug-guix@gnu.org)
5cacf4b49b42b6c8f29ea14772ac4068@lepiller.eu
Hi,
I've updated to the latest guix and upgraded my neovim package. Unfortunately, it segfaults:
$ nvimErreur de segmentation (core dumped)
Here is my current guix:
Génération 34 22 févr. 2019 14:01:39 (actuelle) guix 28cbf65 URL du dépôt : https://git.savannah.gnu.org/git/guix.git branche: master commit : 28cbf65cb944606aafc8b07edd2c521ca58c3127
GUIX_PACKAGE_PATH="/home/tyreunom/nosave/guix-more"
I'm not sure how to debug that, so I'd appreciate some help here :)
R
R
Ricardo Wurmus wrote on 22 Feb 2019 16:06
(name . Julien Lepiller)(address . julien@lepiller.eu)(address . 34616@debbugs.gnu.org)
877edr50co.fsf@elephly.net
Julien Lepiller <julien@lepiller.eu> writes:
Toggle quote (20 lines)> Hi,>> I've updated to the latest guix and upgraded my neovim> package. Unfortunately, it segfaults:>> $ nvim> Erreur de segmentation (core dumped)>> Here is my current guix:>> Génération 34 22 févr. 2019 14:01:39 (actuelle)> guix 28cbf65> URL du dépôt : https://git.savannah.gnu.org/git/guix.git> branche: master> commit : 28cbf65cb944606aafc8b07edd2c521ca58c3127>> GUIX_PACKAGE_PATH="/home/tyreunom/nosave/guix-more">> I'm not sure how to debug that, so I'd appreciate some help here :)
I cannot reproduce this with
./pre-inst-env guix environment --pure --ad-hoc neovim
on the same commit.
This is on a foreign distro.
-- Ricardo
J
J
Julien Lepiller wrote on 22 Feb 2019 16:13
(address . 34616@debbugs.gnu.org)
04f4ff1028b0defe4de7d72236f07a90@lepiller.eu
Le 2019-02-22 16:06, Ricardo Wurmus a écrit :
Toggle quote (30 lines)> Julien Lepiller <julien@lepiller.eu> writes:> >> Hi,>> >> I've updated to the latest guix and upgraded my neovim>> package. Unfortunately, it segfaults:>> >> $ nvim>> Erreur de segmentation (core dumped)>> >> Here is my current guix:>> >> Génération 34 22 févr. 2019 14:01:39 (actuelle)>> guix 28cbf65>> URL du dépôt : https://git.savannah.gnu.org/git/guix.git>> branche: master>> commit : 28cbf65cb944606aafc8b07edd2c521ca58c3127>> >> GUIX_PACKAGE_PATH="/home/tyreunom/nosave/guix-more">> >> I'm not sure how to debug that, so I'd appreciate some help here :)> > I cannot reproduce this with> > ./pre-inst-env guix environment --pure --ad-hoc neovim> > on the same commit.> > This is on a foreign distro.
This is where I got neovim from:
téléchargement depuis https://berlin.guixsd.org/nar/gzip/d8ibld5vpsgq7is3k3sf5gqj0i7sgmbh-neovim-0.3.4...
and I can reproduce in the pure environment, on guixsd and on a foreign distro...it seems that GUIX_PACKAGE_PATH is irrelevant, since I can reproduce on a machinethat's on the same commit, without GUIX_PACKAGE_PATH.
I'm on x86_64.
R
R
Ricardo Wurmus wrote on 22 Feb 2019 16:57
(name . Julien Lepiller)(address . julien@lepiller.eu)(address . 34616@debbugs.gnu.org)
8736of4y00.fsf@elephly.net
Julien Lepiller <julien@lepiller.eu> writes:
Toggle quote (5 lines)> This is where I got neovim from:>> téléchargement depuis> https://berlin.guixsd.org/nar/gzip/d8ibld5vpsgq7is3k3sf5gqj0i7sgmbh-neovim-0.3.4...
I built neovim locally. The binary contents differ from the files onberlin. So there are two problems:
- neovim does not build reproducibly- one of the builds segfaults
Have you tried running it in gdb to get a backtrace?
--Ricardo
J
J
Julien Lepiller wrote on 22 Feb 2019 17:35
(address . 34616@debbugs.gnu.org)
767bed0d2f47c74490a878b801dc715d@lepiller.eu
Le 2019-02-22 16:57, Ricardo Wurmus a écrit :
Toggle quote (18 lines)> Julien Lepiller <julien@lepiller.eu> writes:> >> This is where I got neovim from:>> >> téléchargement depuis>> https://berlin.guixsd.org/nar/gzip/d8ibld5vpsgq7is3k3sf5gqj0i7sgmbh-neovim-0.3.4...> > I built neovim locally. The binary contents differ from the files on> berlin. So there are two problems:> > - neovim does not build reproducibly> - one of the builds segfaults> > Have you tried running it in gdb to get a backtrace?> > --> Ricardo
I think that's what you mean:
Thread 2 "nvim" received signal SIGSEGV, Segmentation fault.[Switching to Thread 0x7ffff6206700 (LWP 28664)]0x00007ffff79682ad in __strcmp_avx2 () from /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6(gdb) bt#0 0x00007ffff79682ad in __strcmp_avx2 () from /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6#1 0x0000000000505b31 in strequal ()#2 0x000000000045681e in tui_tk_ti_getstr ()#3 0x00007ffff7e649a5 in try_load_terminfo_key () from /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1#4 0x00007ffff7e64b59 in load_terminfo () from /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1#5 0x00007ffff7e64eda in start_driver () from /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1#6 0x00007ffff7e6007e in termkey_start () from /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1#7 0x0000000000459368 in term_input_init ()#8 0x000000000045951d in tui_main ()#9 0x0000000000449c84 in ui_thread_run ()#10 0x00007ffff7e96019 in start_thread () from /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libpthread.so.0#11 0x00007ffff790c92f in clone () from /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6
R
R
Ricardo Wurmus wrote on 22 Feb 2019 18:11
(name . Julien Lepiller)(address . julien@lepiller.eu)(address . 34616@debbugs.gnu.org)
87r2bz3fzb.fsf@elephly.net
Julien Lepiller <julien@lepiller.eu> writes:
Toggle quote (9 lines)> I think that's what you mean:>> Thread 2 "nvim" received signal SIGSEGV, Segmentation fault.> [Switching to Thread 0x7ffff6206700 (LWP 28664)]> 0x00007ffff79682ad in __strcmp_avx2 () from> /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6> (gdb) bt> #0 0x00007ffff79682ad in __strcmp_avx2 () from
avx2…? Is this indicative of libtermkey being tuned to a certainCPU type?
Toggle quote (12 lines)> /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6> #1 0x0000000000505b31 in strequal ()> #2 0x000000000045681e in tui_tk_ti_getstr ()> #3 0x00007ffff7e649a5 in try_load_terminfo_key () from> /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> #4 0x00007ffff7e64b59 in load_terminfo () from> /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> #5 0x00007ffff7e64eda in start_driver () from> /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> #6 0x00007ffff7e6007e in termkey_start () from> /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1
So the problem is in libtermkey. Can we reproduce this with anotherpackage using libtermkey?
--Ricardo
E
E
Efraim Flashner wrote on 24 Feb 2019 10:02
(name . Ricardo Wurmus)(address . rekado@elephly.net)
20190224090200.GH18280@macbook41
On Fri, Feb 22, 2019 at 06:11:52PM +0100, Ricardo Wurmus wrote:
Toggle quote (16 lines)> > Julien Lepiller <julien@lepiller.eu> writes:> > > I think that's what you mean:> >> > Thread 2 "nvim" received signal SIGSEGV, Segmentation fault.> > [Switching to Thread 0x7ffff6206700 (LWP 28664)]> > 0x00007ffff79682ad in __strcmp_avx2 () from> > /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6> > (gdb) bt> > #0 0x00007ffff79682ad in __strcmp_avx2 () from> > avx2…? Is this indicative of libtermkey being tuned to a certain> CPU type?>
grepping the source doesn't lead me to believe so
Toggle quote (16 lines)> > /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6> > #1 0x0000000000505b31 in strequal ()> > #2 0x000000000045681e in tui_tk_ti_getstr ()> > #3 0x00007ffff7e649a5 in try_load_terminfo_key () from> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> > #4 0x00007ffff7e64b59 in load_terminfo () from> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> > #5 0x00007ffff7e64eda in start_driver () from> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> > #6 0x00007ffff7e6007e in termkey_start () from> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1> > So the problem is in libtermkey. Can we reproduce this with another> package using libtermkey?>
The only other package which uses libtermkey is vis, a text editor.
-- Efraim Flashner <efraim@flashner.co.il> אפרים פלשנרGPG key = A28B F40C 3E55 1372 662D 14F7 41AA E7DC CA3D 8351Confidentiality cannot be guaranteed on emails sent or received unencrypted
-----BEGIN PGP SIGNATURE-----
iQIzBAABCgAdFiEEoov0DD5VE3JmLRT3Qarn3Mo9g1EFAlxyXYgACgkQQarn3Mo9g1G73g//TR4aKRSbEQXTK2O1QXzLp4o5EGwAI5tM8CxpvnLWk6IfEIHArcJ5tluPVrdbBj8kFs23dnKKe7cC4hjL/pPmWUP4P5jMbQrhbnMWBr0P3myODQ81UoIg5XhyNjfptKSdve71wwU1GTZpOVHGJyqPe06iPOabhiBbNgLy0rf53DOlyxkKEcyHosgUicf8rJn91hFanJSQ3cGHu9PlYB5QM65TLCF6KiUfVb4Bfi6CCZEIpUZGiRIK+4QBm0rZwaSRulcaX5u6uzXlRvilEjqxCluYyrR/CU/w4JHDRvdslNVhkneN2WK8a6XQ1WP8zlwSjzGFNmGR+eNqX/VJj28/5WpSmPZqBk+OXQlqDVK642/3DmycAQBx/hCY3Z53JULyaeaX3WYkxT1OeoUOtpbhwURxSuo+buDMpA8SCinkJJUzY8FJwQW7TtEtiFbtgDiFLFa78/kMqE54vLkS1c7r+nqRmLypKd/fdTHThmgU0QqmdnDRK5p+uo/zpOtv0ZjjVuiX1xwYdmbBYSobbqZtBmM2nh1tL6qqdlW4+5QbED7ISmQTTN5Xr+wP5TCXcUKNxkaedzHnue/MOosbghBzw4gbAFfCadl9gPG6Holq9DEl3fwN/7Xy9zqCr2YLkGDSuPHpOzKoqZ7CXSD/dJYAyEuqgIyMq+q4USTVzcReRuw==irWe-----END PGP SIGNATURE-----

R
R
Ricardo Wurmus wrote on 1 Mar 2019 17:31
(name . Efraim Flashner)(address . efraim@flashner.co.il)
87a7iey2sv.fsf@elephly.net
Efraim Flashner <efraim@flashner.co.il> writes:
Toggle quote (18 lines)>> > /gnu/store/h90vnqw0nwd0hhm1l5dgxsdrigddfmq4-glibc-2.28/lib/libc.so.6>> > #1 0x0000000000505b31 in strequal ()>> > #2 0x000000000045681e in tui_tk_ti_getstr ()>> > #3 0x00007ffff7e649a5 in try_load_terminfo_key () from>> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1>> > #4 0x00007ffff7e64b59 in load_terminfo () from>> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1>> > #5 0x00007ffff7e64eda in start_driver () from>> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1>> > #6 0x00007ffff7e6007e in termkey_start () from>> > /gnu/store/pl7nh8chyp0av6lb7qck5n9mvvaz24z5-libtermkey-0.21.1/lib/libtermkey.so.1>>>> So the problem is in libtermkey. Can we reproduce this with another>> package using libtermkey?>>>> The only other package which uses libtermkey is vis, a text editor.
vis does not segfault for me.
I tried running the segfaulting nvim again, but with the TERM variableunset. It does not segfault.
TERM= /gnu/store/d8ibld5vpsgq7is3k3sf5gqj0i7sgmbh-neovim-0.3.4/bin/nvim
Any other value for TERM that I tried leads to a lookup of the terminfofiles provided by the ncurses package and subsequently leads to asegfault.
I would also like to point out that this does not segfault:
guix environment --container --ad-hoc neovim -- nvim
Outside of the container ~/.guix-profile/share/terminfo is available,which resides in my profile because I happen to have rxvt-unicodeinstalled. This package provides these files:
./share/terminfo/r/rxvt-unicode/rxvt-unicode{,-256color}
The value of TERM in my sessions is “xterm-256color”.
--Ricardo
B
B
Bradley Haggerty wrote on 5 Mar 2019 20:53
neovim segfaults
(address . 34616@debbugs.gnu.org)
CABGw91dAgDbZgq3SgmZ4wOXVbwZj-trE4oRO=OsCfv5cQTm8Sg@mail.gmail.com
Jumping in here to share some of my own info as I'm having the same issue.
brad@kazuki:~/ > nvimzsh: segmentation fault nvim
I've got neovim installed in the system profile (so root has a comfortableeditor instead of just nano)guix (GNU Guix) d22d246a256814784dfb03437949bdc2efd746a5brad@kazuki:~/ > echo $TERMscreen-256colorHere is my $TERM. I am running inside tmux using termite as my terminalemulator.For the record I get the same segmentation fault outside of tmux using thesame terminal emulator.brad@kazuki:~/ > echo $TERMxterm-termite
I tried the guix environment --container command that Ricardo shared and inthat instance nvim works for me in tmux just fine as well. If there's moreinfo about my environment I can share to help track down the issue, pleaselet me know.
Attachment: file
G
G
Gábor Boskovits wrote on 7 Mar 2019 11:25
(address . 34616-done@debbugs.gnu.org)
CAE4v=pjEU7bfpLFmZgFxThcexhS8B7tsW2LLCaSrMG7G7mrR_A@mail.gmail.com
This is fixed by0a10abf73e7bb1d5f751229c709903a0f66574d2and2702aa913648b308a4fdac0b2b27b722894265edon master.
Closed
?
Your comment

This issue is archived.

To comment on this conversation send email to 34616@debbugs.gnu.org