[MU3] Solve a crash when score contains a vbox which is not on top of a page. #7778
Conversation
libmscore/layout.cpp
Outdated
| @@ -5120,7 +5120,7 @@ VerticalGapData::VerticalGapData(bool first, System *sys, Staff *st, SysStaff *s | |||
| _normalisedSpacing = system->y() + (sysStaff ? sysStaff->y() : 0.0) - y; | |||
| _maxActualSpacing = system->score()->styleP(Sid::maxStaffSpread); | |||
|
|
|||
| Spacer* spacer { sys->upSpacer(st->idx(), nextSpacer) }; | |||
| Spacer* spacer { st ? sys->upSpacer(st->idx(), nextSpacer) : nullptr }; | |||
There was a problem hiding this comment.
No, there is always a system, a VBox is also a system. However, a VBox doesn't have any SysStaff's and that caused the crash (if the VBox was somewhere in the middle of the page).
There was a problem hiding this comment.
So maybe it'd never happen (sounds like 'famous last words' though), but that check won't harm either
There was a problem hiding this comment.
If a test is required, this will be a wrong place since system is already dereferenced several time times before (lines 5540, 5543 and 5544). Also before calling the constructor the system argument is dereferenced already.
There was a problem hiding this comment.
system is used before, not sys. Except in the constructor, but who knows what system(sys) does if sys is a nullptr.
but yes, probably not need to check, not there at least
anyway: how to reproduce the issue prior to this change?
There was a problem hiding this comment.
This is in the constructor and with system(sys) the member system is initialized with the argument sys, so both are the same.
The issue can be reproduced creating a VBox somewhere after the first system.
There was a problem hiding this comment.
We, no, I can't reproduce the issue.
And a mix of using system and sys looks wrong to me
There was a problem hiding this comment.
You tried to reproduce with 3.6.2 of the 3.x branch? The crash occurs in the 3.x branch (and master) only, not in 3.6.2 .
Strictly speaking, mix of sys and system is not "wrong" but it certainly is confusing and is corrected.
There was a problem hiding this comment.
Ah, I tried to reproduce in 3.6.2, so that crash got to be a later regression.
And yes, 'wrong' is a bit strong, but certainly confusing.
There was a problem hiding this comment.
Yes, a later regression due to PR #7529
njvdberg
force-pushed
the
spacer-crash-3x
branch
from
5c7446e
to
477013f
Compare
When a
VBoxfound, aVerticalGapDataobject is created but if thisVBoxwasn't the first element on a page, the application crashed because the constructor was looking for aSpaceron a non-existing staff.PR #7779 is similar for
master.