Cycle Detection is maybe broken in 2.94 #1064
Labels
No labels
Affects/CppNix
Affects/Nightly
Affects/Only nightly
Affects/Stable
Area/build-packaging
Area/cli
Area/evaluator
Area/fetching
Area/flakes
Area/language
Area/lix ci
Area/nix-eval-jobs
Area/profiles
Area/protocol
Area/releng
Area/remote-builds
Area/repl
Area/repl/debugger
Area/store
awaiting
author
awaiting
contributors
bug
Context
contributors
Context
drive-by
Context
maintainers
Context
RFD
crash 💥
Cross Compilation
devx
docs
Downstream Dependents
E/easy
E/hard
E/help wanted
E/reproducible
E/requires rearchitecture
Feature/S3
imported
Language/Bash
Language/C++
Language/NixLang
Language/Python
Language/Rust
Needs Langver
OS/Linux
OS/macOS
performance
regression
release-blocker
stability
Status
blocked
Status
invalid
Status
postponed
Status
wontfix
testing
testing/flakey
Topic/Large Scale Installations
ux
No project
No assignees
4 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
lix-project/lix#1064
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Describe the bug
In 2.94, something about cyclic dependency detection seems to be broken.
If you add cyclic outputs to libnet, Lix fails with:
Steps To Reproduce
Add this patch to Nixpkgs:
Expected behavior
Lix should show me how this cycle happened.
nix --versionoutputcc @nikstur
Reproduced on
fd38f625ewithnix build --impure -E 'with import <nixpkgs> { }; libnet.overrideAttrs { outputs = [ "out" "dev" ]; }' --no-use-cgroupsDaemon's stack trace:
This doesn't fully happen before
e2641cb890(cc @ma27), but on that commit the daemon doesn't fully crash, merely hanging up early (and no stack trace).Ah, no I think our crash handler is just being unreliable. So: bisected to
e2641cb890, cc @ma27.I'm aware and looking.
OK, the culprit was a very dumb oversight of mine, I didn't notice that this will break once an output not only has references to other outputs, but also to paths from a different derivation (and this only got triggered on this error, so the blast-radius is luckily rather low).
I'm very sorry for this.
Expect a patch today.
EDIT: I think we should ship this in 2.94.1. Feel free to remove again if you disagree.
This issue was mentioned on Gerrit on the following CLs: