Issues

AsRelative / GetRelativePathTo / SemanticRelativePath.Make throw an undocumented "must be relative" ArgumentException when the two paths are on different roots

ktsu-dev/Semantics#527 · opened

on Oct 31, 2026

1

Mixing from a transparent colour darkens the result: Transparent.MixOklab(White, 0.5) is #63636380 because alpha is interpolated straight, not premultiplied

ktsu-dev/Semantics#526 · opened

on Oct 31, 2026

decimal and Half unit factories silently return wrong values when a conversion factor doesn't fit: Mass<decimal>.FromAtomicMassUnit(12) is 2.04e-26 kg (2.4% high), Energy<Half>.FromKilowattHour(1) is Infinity

ktsu-dev/Semantics#525 · opened

on Oct 31, 2026

1

Color.AdjustForContrast lands exactly on the threshold, so about half its results fail the target once written as hex/bytes (#A7E058 on #8815C9 → AA comes back as #A7E058, still 4.486:1)

ktsu-dev/Semantics#522 · opened

on Oct 31, 2026

1

Same-type Cross() on 3D quantities returns the operand's dimension, so Displacement3D × Displacement3D (m²) types as a length

ktsu-dev/Semantics#521 · opened

on Oct 30, 2026

1

StorageMath.Sqrt/Hypot return the wrong floor for long, ulong, Int128 and UInt128 past 2^53, and Hypot throws OverflowException near MaxValue

ktsu-dev/Semantics#520 · opened

on Oct 30, 2026

1

Benchmark history silently loses a release when a later CI run on main finishes while a benchmark run is in progress

ktsu-dev/Semantics#518 · opened

on Oct 30, 2026

1

No release since v5.12.1: every CI run on main fails in Release on package validation (CP0014/CP0015) for Color, Cpp, Music and Strings.Identifiers

ktsu-dev/Semantics#517 · opened

on Oct 30, 2026

"C:" validates as both an absolute and a relative path on Windows, and its Parent/IsRoot/Depth depend on the drive's current directory

ktsu-dev/Semantics#516 · opened

on Oct 30, 2026

1

Benchmark history orders a prerelease after its own release (5.11.1-pre.1 charted as newer than 5.11.1)

ktsu-dev/Semantics#515 · opened

on Oct 30, 2026

1

Strings benchmark's Ordering pair now compares culture-sensitive CompareTo against the ordinal SemanticString.CompareTo, so its 1.11 ratio doesn't measure the wrapper

ktsu-dev/Semantics#514 · opened

on Oct 30, 2026

1

IVector0 and IVector1 declare no arithmetic operators, so generic code over them can't add or scale (strategy doc's interface spec)

ktsu-dev/Semantics#513 · opened

on Oct 30, 2026

1

A minor key's leading-tone chord (G#dim / G#dim7 in A minor) is reported as Chromatic, though V/V7 counts as diatonic and the spec maps vii° to Dominant

ktsu-dev/Semantics#512 · opened

on Oct 30, 2026

1

`new Uuid()`, `new AbsoluteDirectoryPath()` etc. build a semantic string that never passed validation: the public parameterless constructor is the gap #310 left open

ktsu-dev/Semantics#511 · opened

on Oct 30, 2026

1

`Width + Height`, `Weight - Thrust` and `Position3D + Translation3D` don't compile (CS0019): sibling overloads of one base can't be added or subtracted

ktsu-dev/Semantics#510 · opened

on Oct 30, 2026

1

Form never recognises a twelve-bar blues in pentatonic or blues keys (C7–F7–G7 in "C blues_minor" is ThroughComposed)

ktsu-dev/Semantics#508 · opened

on Oct 30, 2026

1

Color.AdjustForContrast inverts a dark foreground to light when a tiny darkening would pass (#303030 on #737373 → #CECECE)

ktsu-dev/Semantics#507 · opened

on Oct 30, 2026

`Position3D.Cross(Force3D)` and `Position3D.Cross(Velocity3D)` don't compile: vector overloads don't forward their base's typed cross products

ktsu-dev/Semantics#506 · opened

on Oct 30, 2026

1

`Mass * Acceleration3D`, `Mass * Velocity3D` and `Capacitance * Voltage` don't compile: products declared on a magnitude-only dimension never reach vector forms

ktsu-dev/Semantics#505 · opened

on Oct 30, 2026

1

Paths: `dir.Parent / dir.Name` names a different directory when the path ends in `..` or `.`, breaking the documented round trip

ktsu-dev/Semantics#503 · opened

on Oct 30, 2026

1

Paths: the parent of a one-segment relative path (or the directory of a bare file name) is "", so AsAbsolute() throws and IsDirectory is false

ktsu-dev/Semantics#502 · opened

on Oct 30, 2026

1

NormalizedParameter<int> compiles but Normalize only returns 0 or 1 and Denormalize only reaches Min or Max

ktsu-dev/Semantics#499 · opened

on Oct 29, 2026

2

C++ prelude: `1 / seconds` and `1.0 / seconds` don't compile for the default float representation, and mixed-representation `*` and `/` fail although `+` works since #451

ktsu-dev/Semantics#491 · opened

on Oct 28, 2026

1

ChromaticChords reports a pentatonic key's own ii, IV and V as "borrowed from parallel major", contradicting RomanNumerals/Functions/Cadences (#470 not applied to chromatic analysis)

ktsu-dev/Semantics#490 · opened

on Oct 28, 2026

1

PH.ToPOH() returns a PH<T>, so the pOH of an acid reports IsBasic == true and prints as "pH 11"

ktsu-dev/Semantics#482 · opened

on Oct 20, 2026

1

Navigation Menu

Repository navigation

Issues

AsRelative / GetRelativePathTo / SemanticRelativePath.Make throw an undocumented "must be relative" ArgumentException when the two paths are on different roots

Mixing from a transparent colour darkens the result: Transparent.MixOklab(White, 0.5) is #63636380 because alpha is interpolated straight, not premultiplied

decimal and Half unit factories silently return wrong values when a conversion factor doesn't fit: Mass<decimal>.FromAtomicMassUnit(12) is 2.04e-26 kg (2.4% high), Energy<Half>.FromKilowattHour(1) is Infinity

Color.AdjustForContrast lands exactly on the threshold, so about half its results fail the target once written as hex/bytes (#A7E058 on #8815C9 → AA comes back as #A7E058, still 4.486:1)

Same-type Cross() on 3D quantities returns the operand's dimension, so Displacement3D × Displacement3D (m²) types as a length

StorageMath.Sqrt/Hypot return the wrong floor for long, ulong, Int128 and UInt128 past 2^53, and Hypot throws OverflowException near MaxValue

Benchmark history silently loses a release when a later CI run on main finishes while a benchmark run is in progress

No release since v5.12.1: every CI run on main fails in Release on package validation (CP0014/CP0015) for Color, Cpp, Music and Strings.Identifiers

"C:" validates as both an absolute and a relative path on Windows, and its Parent/IsRoot/Depth depend on the drive's current directory

Benchmark history orders a prerelease after its own release (5.11.1-pre.1 charted as newer than 5.11.1)

Strings benchmark's Ordering pair now compares culture-sensitive CompareTo against the ordinal SemanticString.CompareTo, so its 1.11 ratio doesn't measure the wrapper

IVector0 and IVector1 declare no arithmetic operators, so generic code over them can't add or scale (strategy doc's interface spec)

A minor key's leading-tone chord (G#dim / G#dim7 in A minor) is reported as Chromatic, though V/V7 counts as diatonic and the spec maps vii° to Dominant

`new Uuid()`, `new AbsoluteDirectoryPath()` etc. build a semantic string that never passed validation: the public parameterless constructor is the gap #310 left open

`Width + Height`, `Weight - Thrust` and `Position3D + Translation3D` don't compile (CS0019): sibling overloads of one base can't be added or subtracted

Form never recognises a twelve-bar blues in pentatonic or blues keys (C7–F7–G7 in "C blues_minor" is ThroughComposed)

Color.AdjustForContrast inverts a dark foreground to light when a tiny darkening would pass (#303030 on #737373 → #CECECE)

`Position3D.Cross(Force3D)` and `Position3D.Cross(Velocity3D)` don't compile: vector overloads don't forward their base's typed cross products

`Mass * Acceleration3D`, `Mass * Velocity3D` and `Capacitance * Voltage` don't compile: products declared on a magnitude-only dimension never reach vector forms

Paths: `dir.Parent / dir.Name` names a different directory when the path ends in `..` or `.`, breaking the documented round trip

Paths: the parent of a one-segment relative path (or the directory of a bare file name) is "", so AsAbsolute() throws and IsDirectory is false

NormalizedParameter<int> compiles but Normalize only returns 0 or 1 and Denormalize only reaches Min or Max

C++ prelude: `1 / seconds` and `1.0 / seconds` don't compile for the default float representation, and mixed-representation `*` and `/` fail although `+` works since #451

ChromaticChords reports a pentatonic key's own ii, IV and V as "borrowed from parallel major", contradicting RomanNumerals/Functions/Cadences (#470 not applied to chromatic analysis)

PH.ToPOH() returns a PH<T>, so the pOH of an acid reports IsBasic == true and prints as "pH 11"