fix: preserve cache ID when metadata is content-equivalent #5374
+120
−81
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When the build backend metadata cache is outdated (e.g., due to timestamp changes), pixi would regenerate metadata by calling the build backend again. Previously, this always generated a new random ID, which invalidated downstream caches (like source metadata) even when the actual metadata content was unchanged. This change: - Compares new metadata with the previously cached metadata - Reuses the existing ID if the content is equivalent (excluding id, timestamp, and cache_version fields) - Switches from random u64 IDs to nanoid strings for better uniqueness - Bumps cache version to v1 due to the format change This optimization prevents unnecessary cache invalidation when the build backend returns the same metadata, improving build performance.
baszalmstra
changed the title
baszalmstra
force-pushed
the
claude/fix-metadata-cache-invalidation-5pPf1
branch
2 times, most recently
from
c68717b to
b480472
Compare
This refactoring improves maintainability of the metadata comparison logic: - Created `CachedCondaMetadataContent` struct containing all fields that represent the actual build backend response (project_model_hash, build_source, build_variants, build_variant_files, input_globs, input_files, outputs) - `CachedCondaMetadata` now wraps this content struct with cache metadata (id, cache_version, timestamp) - Changed `BinaryHeap` to `BTreeSet` for build_variant_files and input_globs to enable `PartialEq`/`Eq` derives - Simplified `is_content_equivalent()` to just compare `self.content == other.content` - Refactored `verify_cache_freshness` to take a reference and return bool, avoiding unnecessary clones when cache is stale This makes the equality comparison less error-prone: when fields are added or removed from `CachedCondaMetadataContent`, the derived `PartialEq` automatically includes them, preventing bugs from manual field listing.
baszalmstra
force-pushed
the
claude/fix-metadata-cache-invalidation-5pPf1
branch
from
b480472 to
74ba0c7
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
When the build backend metadata cache is outdated (e.g., due to timestampchanges), pixi would regenerate metadata by calling the build backend again. Previously, this always generated a new random ID, which invalidated downstream caches (like source metadata) even when the actual metadata content was unchanged.
This change:
This optimization prevents unnecessary cache invalidation when the build backend returns the same metadata, improving build performance.
How Has This Been Tested?
AI Disclosure
Tools: Claude Code Opus 4.5
Checklist:
schema/model.py.