It's hard to track where a source came from inside the evaluator #669
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
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
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 milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: lix-project/lix#669
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?
Raised from #668: if you have some store path with busted nix code in it, it's really hard to find the code that fetched it to figure out why it even got evaluated. We can surely store that information in the evaluator and surface it somehow (maybe in stack traces?).
This is especially true of flakes where the path is just "source", which is useless. It should give the attr path from which it came.
the issue in #668 was even worse since the problematic code came in via an imperative nixpkgs overlay. provenance tracking that can detangle this mess is going to be extremely costly :( probably not a large issue is we can make it free unless a special flag is given, but oof.
I think theoretically we could track it as a map from each individual store path going into the evaluation to the positions it's from, or so. So the cost would be just related to the total number of store paths in the evaluation and then the usage side is only when there's failures or debugging happening.
with @piegames's position tracking improvements that haven't landed yet we'll also have the position of the path expression in source. if we were to just add an error frame to import and show two innermost frames stack frames instead of just one we could have something like
instead of
Related: #625