Handle (new) buffer protocol conforming types in Pickle.decode
#143
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
On Python 3, `pickle.loads` is able to work with anything that supports the (new) buffer protocol. Unfortunately Python 2's, `cPickle.loads` is not so flexible and requires a `bytes` object specifically. So this ensures that other objects are coerced to `bytes` objects specifically on Python 2.
On Python 3, `pickle.loads` expects an object that implements the buffer protocol and is contiguous. While it would already raise this error, it helpful for us to check this first. Plus we should be able to handle Fortran order buffers should that be needed. On Python 2, the more strict constraint of a `bytes` object is required, which is still enforced.
This ensures that Pickle can decode anything that supports the (new) buffer protocol.
jakirkham
commented
Nov 30, 2018
| if PY2: | ||
| buf = ensure_bytes(buf) | ||
| else: | ||
| buf = ensure_contiguous_ndarray(buf) |
Member
Author
There was a problem hiding this comment.
Just to provide context from the Python codebase, Python 2 requires bytes here and Python 3 uses the (new) buffer protocol to get a C-contiguous array.
jakirkham
commented
Nov 30, 2018
|
|
||
| enc = codec.encode(arr) | ||
| dec = codec.decode(ensure_ndarray(enc)) | ||
| assert_array_items_equal(arr, dec) |
Member
Author
There was a problem hiding this comment.
Seems this worked for everything except Pickle on Python 2 as demonstrated by this CI failure. This is likely a consequence of the improved buffer handling under the hood, which MsgPack, JSON, etc. all use already. Pickle was the only one not using these functions, which this PR fixes. So testing this generally seems to be fine. All the other codecs already test decoding with ndarray.
7 tasks
Member
Author
|
Thanks for taking a look. |
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.
Python 3's
pickle.loadsis able to handle anything that conforms to the (new) buffer protocol as long as it is C-contiguous. We go ahead and enforce this on all inputs; raising if that is not possible. On Python 2cPickle.loadsis not as flexible as it requires abytesobject explicitly. This enforces that case as well.TODO:
tox -e py37passes locallytox -e py27passes locallytox -e docspasses locally