[hjmjohnson]

ImageRegionReverseConstIterator sets its row span one pixel too far when placed with SetIndex() or converted from another iterator, so operator++ can walk outside a cropped region. This computes both span offsets in one helper that matches GoToBegin(), and adds a test.

Found while triaging #7065, where the same formula moved into InitializeSpans(). The bug is on main independently of that PR, so the fix is kept separate to keep #7065 small.

Reproduction and test

The test uses a 10 x 4 image and a region starting at (2, 1) with size (3, 2). It places a reverse iterator at (4, 2), the row's last pixel, and at (3, 2), mid-row, and steps each to the end. Both placements fail on main.

Path	Visited on main
SetIndex((4, 2))	(4,2) (3,2) (2,2) (1,2) (0,2) (4,1) (3,1)
converted from an ImageRegionIterator at (4, 2)	(4,2) (3,2) (2,2) (1,2) (4,1) (3,1) (2,1)
expected, both paths	(4,2) (3,2) (2,2) (4,1) (3,1) (2,1)

The span began one past the row's last pixel. That put the span end on the row's first pixel instead of one before it. The converting constructors also computed m_SpanEndOffset in the member initializer list from a m_SpanBeginOffset that had not been assigned yet.

With the fix, both paths match the expected order. All 723 ITKCommon tests pass on macOS arm64 (clang 19, Release).

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no blocking issues were found.

Summary

Fixes reverse region iterators that leave a cropped region after placement or conversion.

• Reverse region iterators stay inside cropped rows after placement.

_{Reviews (2) · Last reviewed commit: "BUG: Keep reverse region iterators insid..." · Reviewed by Greptile}

[dzenanz]

Change looks good on a glance.

[hjmjohnson]

@greptileai review