My previous patch (02-fix_set_item_calls.dpatch) for this problem fixed the FTBFS but included a subtle regression (thanks James Westby for noticing).
This new patch (02-fix-ftbfs-and-hrdiagram-opts.dpatch) REPLACES the previous patch. It fixes the problem, and has been more carefully tested -- it fixes the regression and a subsequent race condition bug, both necessary to implement the intended functionality. The subtle regression: The checks in set_{label,diameter}_item e.g. if (globals::chart_rules.StarLabels == l) return; need to be pushed down a few lines -- past the program_hr stuff. The race condition bug: The set_{label,diameter}_item() accessor routines were checking globals::hr_program_viewer to determine whether the "HR Diagram" window had been started or not -- but that global doesn't get set until the configure event (maybe a race condition even, for other call stacks). The set routines need to check globals::hr_program_window instead, and the hr_quit() routine needs to stomp that global when the window goes down.
02-fix-ftbfs-and-hrdiagram-opts.dpatch
Description: application/shellscript