Help with corrupted measures in 1.3 file

See attached file “SCHÖN ToH Snippet 1.3 Pf #100- Henjor CORRUPT m479.mscz” (from thread Schoenberg's Theory), measure 479 (#141/b, 5th measure; cf page 263 of pdf here). Deleting the mess, even deleting every corrupt measure and the following blank ones as well as several preceding ones, appending new blanks and renotating them does not help: Notes look fine when doing it, saving also, but reload the file and – voilà, back again! Go to measure 491, select 3rd note (C in voice 1): Playback has 3 simultaneous notes, all in voice 1. Select any of the 3, playback has them all. (They are properly notes 5, 8 & 9 in a phrase.) Curiously, #142 a) looks OK but is not (see playback). Also, vertically crammed manually adjusted spacing is no problem (see measures 269—84, —331). Have done many partly remakes and several full ones with different deletions, it repeats exactly and I cannot proceed.

Digging around, I have only managed to vary the corruption, occurring some measures earlier. (See attached file “Sch ToH #(100-)141 Henjor CORRUPT m470.mscz”.)

Opened in nightly: select first corrupt measure, open in debugger. LayoutBreak is OK, but when selecting first Segment causes nightly.exe to crash:
“This application has requested the Runtime to terminate in an unusual way.
Please contact the application's support team for more information.”
Clicking certain objects in the affected measures also causes instant crash. (Tested most recently with MuseScoreNightly-2015-02-15-0807-02a3710.)

I have the impression that the problem comes from something before the awkward notation of measures 448—74 (#141/a) – maybe even before measures 389—95 (#138) – but that corruption is triggered by saving something added in a voice > 1, then file reload. (I think I once saw rests alone jumping left upon save, then whole measure corruption only at file reload.)
(#138 is done in 2 voices; #141/a must have 5, 1 of which lifted from the lower stave.)

Reasons for doing this in 1.x are in the above-mentioned thread, but the file is intended for later 2.x update. I need a ‘healthy’ file, if possible also to know what not to do and whether to do the risky portion in 2.x instead.

Suggestions?