Since the overlap test keeps integer sub-cells of an accepted cell (9b4acd6ab), a half-volume primitive
cell whose metric is C-centred monoclinic reaches the candidate selection. Its conventional C cell is
the true primitive cell, and IndexedFraction counted spots on that conventional cell ignoring the
centring, so it indexed exactly what the P cell indexes; the sub-cell rule, comparing primitive
volumes, then took the C label. On a monoclinic P crystal with no centring pseudo-symmetry (h+k odd as
strong as h+k even at every resolution) the first pass came out "monoclinic C", and with a
user-fixed P 1 21 1 the run was refused by the fixed-group lattice check.
A spot now counts only where its rounded conventional index is a node of the centred lattice (an
integer index on the primitive cell). A genuinely centred crystal is unaffected: its spots obey the
centring. Measured: 14 witness sets (I, F, R, C lattices, small molecules) give the same first-pass
lattices and merging results; the affected set merges in P 1 21 1 with and without -S.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HVeDQsU5a8AJUdJPgWevAK
leonarski_fandClaude Opus 5.5
35f49a295d