Fix a mis-parse of NPY v2.0 and 3.0 headers #77
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
CNPY originally only supported NPY 1.0 headers, with a 2-byte length field. This commit causes parse_npy_header to respond to NPY 2.0 and 3.0 by reading the header length as 4 bytes.
|
I only saw version 2.0 from NEP 1, can you tell me where is version 3.0 defined? Thanks. |
Author
|
You can find all 3 versions defined on numpy.org. Specifically, on the numpy.lib.format page. Version 3.0 merely swaps ASCII strings to Unicode strings. |
s-trinh
reviewed
Sep 2, 2025
cnpy.cpp
Outdated
| header_len = *reinterpret_cast<uint32_t*>(buffer+8); | ||
| else | ||
| header_len = *reinterpret_cast<uint16_t*>(buffer+8); | ||
| std::string header(reinterpret_cast<char*>(buffer+(extended_header ? 11 : 9)),header_len); |
There was a problem hiding this comment.
Since header_len is uint16_t, I think it should be?
Suggested change
| std::string header(reinterpret_cast<char*>(buffer+(extended_header ? 11 : 9)),header_len); | |
| std::string header(reinterpret_cast<char*>(buffer+(extended_header ? 12 : 10)),header_len); |
Author
There was a problem hiding this comment.
I think you're right. It's been working because the only use of header has been based on relative locations discovered using find and similar, but we could improve performance (ever so slightly) by using the correct offset.
Author
There was a problem hiding this comment.
Fixed. Thanks for pointing this out.
The `header` string was started too early, and had included the final byte of the length field (and omitting the final byte). This commit adjusts where it starts to fix the issue. Note: No ill effects occurred earlier because each parsed field is located using `find`, and the final byte is typically `}` and irrelevant to the parsed fields.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
CNPY originally only supported NPY 1.0 headers, with a 2-byte length field. This PR causes
parse_npy_header()to respond to NPY 2.0 and 3.0 by reading the header length as 4 bytes.