RE-26 (PR #237) measured that refs1[1] of a wall record's +0x88 reference list equals the assigned IfcWallType.Tag on 356/360 walls (newest frame). The four misses are all the 18" Basement type walls: the slot names 1851 while the export's IfcWallType.Tag is 3897 for all four. Both ids are stable across frames, so this is not a stale-frame artifact — it looks like two id spaces for the same type (element id vs symbol id, or a type-graph indirection 1851 -> 3897).
Resolving the mapping would make the type join 360/360 and may explain other +0x88 slots (#228). The thickness join itself is unaffected (the 18" nominal matched via the type's own record).
Measured on 2024_Core_Interior.rvt (sha256 c805df44...). Follow-up from PR #237.
RE-26 (PR #237) measured that refs1[1] of a wall record's +0x88 reference list equals the assigned IfcWallType.Tag on 356/360 walls (newest frame). The four misses are all the 18" Basement type walls: the slot names 1851 while the export's IfcWallType.Tag is 3897 for all four. Both ids are stable across frames, so this is not a stale-frame artifact — it looks like two id spaces for the same type (element id vs symbol id, or a type-graph indirection 1851 -> 3897).
Resolving the mapping would make the type join 360/360 and may explain other +0x88 slots (#228). The thickness join itself is unaffected (the 18" nominal matched via the type's own record).
Measured on 2024_Core_Interior.rvt (sha256 c805df44...). Follow-up from PR #237.