fix: Making Wkb::buf() return slices containing only the WKB and trim the trailing data#88
Conversation
Kontinuation
force-pushed
the
refactor-buf
branch
2 times, most recently
from
32a6552 to
598f9d8
Compare
Kontinuation
force-pushed
the
refactor-buf
branch
from
598f9d8 to
63e52df
Compare
| /// See page 65 of <https://portal.ogc.org/files/?artifact_id=25355>. | ||
| #[derive(Debug, Clone, Copy)] | ||
| pub struct Coord<'a> { | ||
| /// The underlying WKB buffer |
There was a problem hiding this comment.
It would be great to add comments here saying that this buffer is solely the length of one coordinate
There was a problem hiding this comment.
Actually, we do not have this requirement for Coord here. The Coords are constructed using buffers with trailing data, and slicing is only handled in Coord::coord_slice.
I can also slice buf when constructing the Coord, which will make Coord.buf consistent with other types: buf contains only the data needed by that object.
| impl<'a> Coord<'a> { | ||
| pub(crate) fn new(buf: &'a [u8], byte_order: Endianness, offset: u64, dim: Dimension) -> Self { | ||
| pub(crate) fn new(buf: &'a [u8], byte_order: Endianness, dim: Dimension) -> Self { | ||
| Self { |
There was a problem hiding this comment.
Perhaps it would be descriptive to add a
debug_assert_eq!(buf.length(), dim.size() * F64_WIDTH)in this new()?
🤷♂️
|
Thanks! |
|
Awesome! |
|
@kylebarron Hi Kyle, can you create a new release? We need this to make a new release of SedonaDB. Thank you! |
|
I've been busy with conferences and vacation; I just published 0.9.2 |
4 participants
CHANGES.mdif knowledge of this change could be valuable to users.Fix #86
This is a follow-up of #87. We refactored the code to remove the
buffield fromWkband addbufmethods for geometry types to return sliced WKB buffers. There are also some refactorings to simplify various offset computations included in this PR.