News:

Simutrans.com Portal
Our Simutrans site. You can find everything about Simutrans from here.

[r8296] dragging signal tool causes SIGSEGV

Started by captain crunch, September 27, 2017, 01:16:09 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

captain crunch

Dragging the rail signal tool causes the program to terminate with SIGSEGV.

gdb backtrace:

Core was generated by `./sim -objects pak'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x000000000071c9f1 in vector_tpl<unsigned char>::append (this=0x3e9a5f0, elem=@0x7ffcf850491f: 8 '\b')
    at dataobj/../tpl/vector_tpl.h:96
96                              data[count++] = elem;
[Current thread is 1 (Thread 0x7f2de2fdd900 (LWP 31593))]
(gdb) bt
#0  0x000000000071c9f1 in vector_tpl<unsigned char>::append (this=0x3e9a5f0, elem=@0x7ffcf850491f: 8 '\b')
    at dataobj/../tpl/vector_tpl.h:96
#1  0x000000000070ea02 in tool_build_roadsign_t::mark_tiles (this=0x3e9a550, player=0x36df2b0, start=..., ziel=...)
    at simtool.cc:4839
#2  0x00000000006efab9 in two_click_tool_t::move (this=0x3e9a550, player=0x36df2b0, buttonstate=<optimized out>,
    pos=...) at simmenu.cc:1055
#3  0x00000000006dd0c7 in interaction_t::move_cursor (this=this@entry=0x4096be0, ev=...) at siminteraction.cc:106
#4  0x00000000006ddb54 in interaction_t::process_event (this=this@entry=0x4096be0, ev=...) at siminteraction.cc:412
#5  0x00000000006ddf43 in interaction_t::check_events (this=0x4096be0) at siminteraction.cc:439
#6  0x000000000073c204 in karte_t::interactive (this=0x41d11b0, quit_month=2147483647) at simworld.cc:6580
#7  0x00000000006e6a46 in simu_main (argc=3, argv=0x7ffcf8506c18) at simmain.cc:1348
#8  0x00000000006f9726 in sysmain (argc=3, argv=0x7ffcf8506c18) at simsys.cc:824
#9  0x00007f2de1983b45 in __libc_start_main (main=0x40a7a0 <main(int, char**)>, argc=3, argv=0x7ffcf8506c18,
    init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7ffcf8506c08)
    at libc-start.c:287
#10 0x000000000040a7ce in _start ()


edit: fixed typo

Dwachs

#1
As the screenshot suggests, this crash happens also in very simple situations? Will try to reproduce in the next days

Edit: I cannot reproduce this.
Parsley, sage, rosemary, and maggikraut.

Ters

The debug information suggests that this is from some 64-bit build. As far as I know, that is pretty much unsupported. Unsupported in what I believe is the original sense: we won't stop you, but you get to deal with stuff yourself. (Which I did years ago, but not anymore.)

captain crunch

Please apologize me, it seems the PEBKAC. After doing a 'make clean' and a new build the problem vanished into thin air.

Ters

The problem might still be there, it is just that the rebuild caused a different alignment of the code. A new rebuild could make it appear again.

But it is indeed also possible that some out-of-date files went into the linking process. That would be a bug in itself, except in the build scripts rather than the source code.