# Segfault in nix eval when requesting existant output

**URL:** <https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730>\
**Category:** Bug Reports\
**Created:** [April 18, 2023, 9:21pm UTC](https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730 "2023-04-18T21:21:47Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![jganong](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.flox.dev/jganong/32/152_2.png) [@jganong](https://discourse.flox.dev/u/jganong)\
**Post date:** [April 18, 2023, 9:21pm UTC](https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730/1 "2023-04-18T21:21:47Z")

</div>

‘’’  
[default] jeg@Jamess-MacBook-Air aniMotum\_nix % nix eval ‘nixpkgs#libtiff.out’  
zsh: segmentation fault nix eval ‘nixpkgs#libtiff.out’  
‘’’

what i meant to type was:  
nix eval ‘nixpkgs#libtiff.outputs’

out is one of the outputs of libtff.

if you instead do i non-existant output, it prints an error message.

[default] jeg@Jamess-MacBook-Air aniMotum\_nix % nix eval ‘nixpkgs#libtiff.foobar’  
error: flake ‘flake:nixpkgs’ does not provide attribute ‘packages.aarch64-darwin.libtiff.foobar’, ‘legacyPackages.aarch64-darwin.libtiff.foobar’ or ‘libtiff.foobar’

but if the output does exist (for example “out” is listed by .outputs), it segfaults:

[default] jeg@Jamess-MacBook-Air aniMotum\_nix % nix eval ‘nixpkgs#libtiff.out’  
zsh: segmentation fault nix eval ‘nixpkgs#libtiff.out’

this is not a horrible bug to me, i just mention in case it you want to fix it.

---

<div class="post-metadata">

**Author:** ![aameen](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.flox.dev/aameen/32/564_2.png) [@aameen](https://discourse.flox.dev/u/aameen)\
**Post date:** [April 19, 2023, 3:54pm UTC](https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730/2 "2023-04-19T15:54:23Z")

</div>

Thanks for the report, I’ll get this filed and prioritized.

---

<div class="post-metadata">

**Author:** ![aameen](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.flox.dev/aameen/32/564_2.png) [@aameen](https://discourse.flox.dev/u/aameen)\
**Post date:** [May 10, 2023, 5:19pm UTC](https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730/3 "2023-05-10T17:19:04Z")

</div>

It looks to me like this issue has been fixed by `nix` upstream.  
I also want to highlight that since this wasn’t an error with `flox` itself, this bug is more appropriately reported to [GitHub - NixOS/nix: Nix, the purely functional package manager](https://github.com/NixOS/nix).

---

<div class="post-metadata">

**Author:** ![jganong](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.flox.dev/jganong/32/152_2.png) [@jganong](https://discourse.flox.dev/u/jganong)\
**Post date:** [May 10, 2023, 6:20pm UTC](https://discourse.flox.dev/t/segfault-in-nix-eval-when-requesting-existant-output/730/4 "2023-05-10T18:20:35Z")

</div>

thanks!

i am working on a big project for my lab right now,

and it just uses plain old matlab,

so my flox/nix adventures are on pause.

before my pause, i spoke with the hpc people at my university (we call it the sherlock cluster here)

and they said that they mainly like singularity for this kind of task. i think i can generate a container

with flox and then run that with singularity. so when i come off my matlab binge that is the

direction i will pursue.
