Only mark unaligned types in buffers #505
Merged
+13
−9
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
Member
|
cc @aldanor |
aldanor
reviewed
Nov 21, 2016
include/pybind11/numpy.h
Outdated
| for (auto& field : ordered_fields) { | ||
| if (field.offset > offset) | ||
| oss << (field.offset - offset) << 'x'; | ||
| // note that '=' is required to cover the case of unaligned fields |
Member
There was a problem hiding this comment.
Minor, but maybe this comment line should be updated to reflect what's happening below
Member
|
This looks ok, I guess, and doesn't seem to break anything (I don't see much value since comparing format strings directly is a bad idea, but it doesn't harm to have aligned fields mark as aligned). |
dean0x7d
reviewed
Nov 21, 2016
include/pybind11/numpy.h
Outdated
| #define PYBIND11_FIELD_DESCRIPTOR_EX(T, Field, Name) \ | ||
| ::pybind11::detail::field_descriptor { \ | ||
| Name, offsetof(T, Field), sizeof(decltype(std::declval<T>().Field)), \ | ||
| std::alignment_of<decltype(std::declval<T>().Field)>::value, \ |
Member
There was a problem hiding this comment.
Any reason not to use the alignof keyword? (AFAIK the template version only exists for historical reasons.)
Minor formatting note: see the position of the continuation \ in comparison to the lines above and below.
Previously all types are marked unaligned in buffer format strings, now we test for alignment before adding the '=' marker.
patstew
force-pushed
the
buffer_alignment
branch
from
25eb980 to
2f197c9
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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Outdated suggestions cannot be applied.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.
Previously all types are marked unaligned in buffer format strings, now we test for alignment before adding the '=' marker.
It's not a problem to mark all types as unaligned per-se, it's equivalent because of the explicit padding, but this goes partway to allowing comparison to format strings from other sources, e.g. numpy. It's much easier to test for alignment here rather than parsing the string and trying to work it out. In combination with my other PR #488 , it will be possible to construct a
vector<struct>from a structured array without casting in some cases (namely where the member names are identical and there's no multibyte padding).