BUG: Fix float-precision division in BresenhamLine::BuildLine

[hjmjohnson]

Fix float-precision division in BuildLine(Direction, length). The expression IdentifierType / float was evaluating in float precision despite storing the result in a double. Adds explicit static_cast<double>() on both operands.

Companion to the release-5.4 backport in #6054 where this was caught by Greptile review.

[greptile-apps]

Greptile Summary

This PR fixes a floating-point precision bug in BresenhamLine::BuildLine(Direction, length) where the division IdentifierType / float (from LType = Vector<float>) was evaluated in float precision before being stored in a double. The fix adds explicit static_cast<double>() on both operands, ensuring the division is performed at double precision.

Confidence Score: 5/5

• This PR is safe to merge — it is a minimal, correct precision fix with no side effects.
• The change is a single targeted cast on both sides of one division. LType = Vector<float> confirms the original operand types (IdentifierType / float), so the fix is correct and complete. No other code paths are affected. All remaining considerations are informational only (no P0/P1 findings).
• No files require special attention.

Important Files Changed

Filename	Overview
Modules/Core/Common/include/itkBresenhamLine.hxx	Adds explicit static_cast<double>() on both operands of the division to ensure double-precision arithmetic instead of float-precision. LType = Vector<float>, so without the cast the division was IdentifierType / float → float, then implicitly widened to double with precision already lost.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["mainDirectionLen\n(IdentifierType / integer)"]
    B["Direction[maxDistanceDimension]\n(float, from LType = Vector<float>)"]
    C["itk::Math::Absolute(...)"]

    subgraph before["Before fix (float precision)"]
        D["integer ÷ float\n→ float result"]
        E["Implicit widen to double\n(precision already lost)"]
        D --> E
    end

    subgraph after["After fix (double precision)"]
        F["static_cast<double>(mainDirectionLen)"]
        G["static_cast<double>(Absolute(...))"]
        H["double ÷ double\n→ double result (full precision)"]
        F --> H
        G --> H
    end

    A --> D
    B --> C --> D
    A --> F
    B --> C --> G

    E --> R1["euclideanLineLen (degraded)"]
    H --> R2["euclideanLineLen (correct)"]

Loading

_{Reviews (1): Last reviewed commit: "BUG: Fix float-precision division in Bre..." | Re-trigger Greptile}