Improve stride copying in generalize a bit #288
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Outdated suggestions cannot be applied.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.
While still not perfect - that should be solved in upstream ndarray, the
new version is an improvement since it checks the compatibility before
attempting to construct the array.
.into_raw_vec()
is very hard to use correctly (this is inherent, it'san access to the raw memory model), so the new version will actually
fail to convert in some cases that were passing as silent errors before
(related to internally sliced arrays, rather uncommon).