Conversation
…ribute updates, cancel timed-out sync coroutines - `AsyncGroup.members`, `Group.members` and `_iter_members_deep` now honor and forward `use_consolidated_for_children`, instead of unconditionally short-circuiting to consolidated metadata when the root has it. - `Group.update_attributes_async` merges new attributes into the existing ones, matching `Group.update_attributes` and `AsyncGroup.update_attributes`. - `sync()` cancels the pending future when it times out instead of leaving the coroutine running on the IO thread. - The `zarr.storage.default_compressor` deprecation warning no longer points at the removed `array.v2_default_compressor` config key. - GPU `NDBuffer.create` zero-fills when no fill value is given, matching the CPU buffer. Addresses B11, B12, B14, B15 and B16 of #181. B13 (`require_array(dtype=None)`) was dropped: zarr-developers#4189 deliberately kept the float64 default. Assisted-by: ClaudeCode:claude-fable-5-1 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #4343 +/- ##
=======================================
Coverage 94.34% 94.34%
=======================================
Files 92 92
Lines 12935 12941 +6
=======================================
+ Hits 12203 12209 +6
Misses 732 732
🚀 New features to boost your workflow:
|
Also specify Zarr format 2 in the deprecated compressor migration example. Assisted-by: Codex:GPT-6
…xercise the recursion fix Address roborev findings on eff4c60: - Rebuild the members-recursion test so a nested group really carries stale consolidated metadata (consolidate at a/b, then create a/b/c). The previous version passed with the `_iter_members_deep` change reverted. - Rename the changelog fragment to the upstream PR number (4343). - Narrow the `update_attributes_async` docstring: merge semantics match `Group.update_attributes`, but it returns a new `Group`. - Reword the GPU zero-fill comment and assert zero-fill in the existing GPU buffer test. Assisted-by: ClaudeCode:claude-fable-5-1 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
🤖 AI text below 🤖
Changes Group.members and AsyncGroup.members so use_consolidated_for_children=False bypasses loaded root consolidated metadata for that traversal, and forwards the option recursively and through the synchronous wrapper. This deliberately changes the previous documented behavior that a loaded root snapshot was always authoritative. It is distinct from the current d-v-b#190 branch, which preserves that behavior.
Also merges attributes in Group.update_attributes_async (returning a new group), requests cancellation after sync timeouts, corrects the default_compressor deprecation example to use a V2 array, and zero-fills GPU NDBuffer.create when fill_value=None to match the CPU implementation.
The require_array dtype change from the original prototype is absent. Cancellation is cooperative: requesting it cannot guarantee that arbitrary coroutine code stops immediately. GPU behavior still requires GPU execution to validate. Historical test counts in the previous description were not rerun during this factual audit.