unstable: fix subsliceOffset regression for non-suffix subslices#1055
Closed
alliasgher wants to merge 3 commits intopelletier:v2from
Closed
unstable: fix subsliceOffset regression for non-suffix subslices#1055alliasgher wants to merge 3 commits intopelletier:v2from
alliasgher wants to merge 3 commits intopelletier:v2from
Conversation
isEmptyStruct returns true whenever a struct has no exported fields, even if that struct holds a meaningful value. This caused types like netip.Addr — which have no exported fields but implement IsZero() and encoding.TextMarshaler — to be unconditionally omitted when tagged with `omitempty`. Fix shouldOmitEmpty by preferring the IsZero() method over the field-scanning fallback, mirroring the logic already used in shouldOmitZero. Additionally, for structs that implement TextMarshaler without an IsZero() method, fall back to reflect.Value.IsZero() instead of the field-based scan. Fixes pelletier#955 Signed-off-by: alliasgher <alliasgher123@gmail.com>
The subsliceOffset helper computed byte offsets by subtracting the length of the slice from the total data length. This only works for tail (suffix) slices; any sub-slice taken from the middle — such as b[0:1] in parseSimpleKey's error path — returns an incorrect offset. This was exposed by pelletier#1041, which correctly fixed a different error position bug, but the existing formula is fundamentally wrong for non-suffix slices. Fix: use pointer arithmetic (via unsafe.Pointer) to compute the actual offset of any subslice relative to p.data, regardless of its length. An empty-slice fallback preserves the existing behaviour for the nil case. Fixes pelletier#1047 Signed-off-by: alliasgher <alliasgher123@gmail.com>
pelletier
requested changes
Apr 16, 2026
Use cap(p.data)-cap(b) to compute the byte offset of a subslice within the parser's data without importing unsafe. This works because cap reports the distance from the slice start to the end of its backing array, so the difference between two subslices of the same array is their offset delta. Also fix errorlint: use errors.As instead of type assertion. Signed-off-by: Ali <alliasgher123@gmail.com>
Owner
|
Merged c171216 to solve this with more testing. I think there is a deeper fix to be made but good enough to fix the symptoms for users. |
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
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.
Fixes #1047.
subsliceOffsetcomputed byte offsets by subtracting the slice length fromlen(p.data). This only works for tail (suffix) slices. Sub-slices taken from the middle — e.g.b[0:1]inparseSimpleKey's error path — return a wrong offset.This was exposed by #1041 (which correctly fixed a different bug): after that change the
parseSimpleKeyerror highlight is stillb[0:1], a non-suffix slice, so the offset computation giveslen(p.data) - 1instead of the actual character position.Fix: use pointer arithmetic (
unsafe.Pointer) to compute the actual offset of any subslice relative top.data. An empty-slice fallback preserves the existing behaviour for nil inputs.Includes
TestRangeOffsetAfterCommentfrom #1047.