🪅 Fix blosc loading #73
Merged
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
Jasper-Bekkers
approved these changes
Dec 8, 2024
Member
That investigation can easily happen async an decoupled from this PR. As the code got fixed before merging. |
jasperdewinther
commented
Dec 8, 2024
src/reader.rs
Outdated
| count: usize, | ||
| ) -> Result<Vec<T>, ParseError> { | ||
| if count == 0 { | ||
| return Ok(vec![T::zeroed(); count]); |
Contributor
Author
There was a problem hiding this comment.
Ya that didn't make sense. I am constructing an empty vec now.
Contributor
Author
|
I'm sorry, at the time of merging I didnt know you were still reviewing. |
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Outdated suggestions cannot be applied.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.
For a while we have not been able to load blosc encoded files. But that time is no more. After a significant debugging session I found out that OpenVDB does hit this check https://github.com/AcademySoftwareFoundation/openvdb/blob/master/openvdb/openvdb/io/Compression.h#L308 and does an early out if count == 0. We didnt do this check resulting in us reading
num_compressed_bytesfrom the reader, reading garbage and offsetting the reader incorrectly.I validated the results against one frame of each of JangaFX's free download models, specifically the ones listed below:
All these models now load correctly. yay.
The main problematic files remaining are specific exports I made myself with Houdini. These models use f32 for the voxel values, unlike Embergen, which uses f16.
This PR fixes #33