Doomsday Engine

Operating System: Linux 2.6.16-gentoo-r13

The segfault occurs sometimes after finishing map 4 of
"The Shores of Hell", while loading map 5; it is hard
to reproduce, but it never occured when loading a map
with setmap; it seems that I have to be in hurry to get
to the next map (i. e. I'm pressing the space bar
repeatedly immediately after the end screen for map 4
appears) to reproduce the bug.

I reproduced the bug in gdb (but I don't know how to
properly debug doomsday since it seems to get all the
user input in the X window system and I've to kill it;
thus, I couldn't backtrack after the segfault occured)
and got this result:

0x080e1ae6 in R_LinkShadow (poly=0x0, subsector=0x0) at
r_shadow.c:279
279 info->shadows = link;

After looking into the code, I wondered how the
arguments of R_LinkShadow could be 0x0, since it's
called (via RIT_ShadowSubsectorLinker) by
P_SubsectorBoxIteratorv and there should be a segfault
before the call if the arguments are 0x0.

Labels: Resources