ALL: Crash on exit
|Reported by:||eriktorbjorn||Owned by:||fingolfin|
This is going to be one annoyingly vague bug report.
I'm filing it partly because I'd like to know if anyone
else has seen anything like it.
I was replaying parts of Loom CD, using an April 5 CVS
snapshot, and on two occasions it crashed for me.
Here's a stack trace from one of the occasions:
#0 Normal3x(unsigned char*, unsigned, unsigned char*,
unsigned char*, unsigned, int, int) (srcPtr=0x404aab88
"#", srcPitch=646, null=0x0,
dstPtr=0xc91da <Address 0xc91da out of bounds>,
height=25) at common/scaler.cpp:792
#1 0x0804dae5 in OSystem_SDL::update_screen()
#2 0x080521ad in Scumm::mainRun() (this=0x816e960) at
#3 0x08052396 in Scumm::go() (this=0x816e960) at
#4 0x080ae56a in main (argc=2, argv=0xbffffbf4) at
The other one was very similar, except I was using a
different scaler at the time.
From what I could tell by using gdb on the core dump,
_hwscreen->pixels was NULL when the scaler function was
called. Which makes no sense at all to me. I thought
the only way to change _hwscreen while the game is
running was to change scalers, and I wasn't at the time.
Of course, I wasn't able to reproduce it once I added
debugging printf()s and assert()s...
Ticket imported from: #716591. Ticket imported from: bugs/735.
Change History (25)
comment:15 by , 16 years ago
|Status:||closed → new|
|Summary:||Mysterious crash in SDL backend → Crash on exit|
comment:17 by , 16 years ago
|Component:||Engine: SCUMM → --Unset--|
|Summary:||Crash on exit → ALL: Crash on exit|