[liquidwar-user] Re: liquidwar race conditions

[Christian Mauduit]

HI,

On Sat, Jun 29, 2002 at 12:57:13PM -0700, Mark Glines wrote:
> Hi!  I'm running into some segfaults when running liquidwar-5.5.2
> on an SMP machine.  Is Allegro fully reentrant?
Err, well, I don't know. I'd say no but I'm not sure, maybe this is
simply a LW bug, and not Allegro's fault...

> One crash happened when repainting that big wobbly mouse pointer
> immediately after a text widget (numeric, with a cursor) received
> a mouseover event on the map selection page (causing that cursor
> to be painted).
Well, to be honest, this screen is programmed in a weird way (translate
-> it's a piece of crap) so I'm not so surprised it crashed.

> Backtraces follow:
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 2051 (LWP 14626)]
> 0x001504d8 in blit () from /usr/lib/liballeg.so.4.0
> (gdb) bt
> #0  0x001504d8 in blit () from /usr/lib/liballeg.so.4.0
> #1  0x00212118 in xdga_extension_name () from /usr/lib/liballeg.so.4.0
> Cannot access memory at address 0x13c
> (gdb) thread 
> [Current thread is 4 (Thread 2051 (LWP 14626))]
> (gdb) thread 1
> [Switching to thread 1 (Thread 1024 (LWP 6106))]#0  0x08050616 in my_edit_proc
>     ()
> (gdb) bt
> #0  0x08050616 in my_edit_proc ()
> (gdb) thread 2
> [Switching to thread 2 (Thread 2049 (LWP 22253))]#0  0x002f0990 in poll ()
>    from /lib/libc.so.6
> (gdb) bt
> #0  0x002f0990 in poll () from /lib/libc.so.6
> #1  0x00349d96 in __pthread_manager () from /lib/libpthread.so.0
> #2  0x00349fc0 in pthread_start_thread () from /lib/libpthread.so.0
> (gdb) thread 3
> [Switching to thread 3 (Thread 1026 (LWP 25915))]#0  0x002f25ae in select ()
>    from /lib/libc.so.6
> (gdb) bt
> #0  0x002f25ae in select () from /lib/libc.so.6
> #1  0x00212118 in xdga_extension_name () from /usr/lib/liballeg.so.4.0
> (gdb) thread 4
> [Switching to thread 4 (Thread 2051 (LWP 14626))]#0  0x001504d8 in blit ()
>    from /usr/lib/liballeg.so.4.0
> (gdb) bt
> #0  0x001504d8 in blit () from /usr/lib/liballeg.so.4.0
> #1  0x00212118 in xdga_extension_name () from /usr/lib/liballeg.so.4.0
> Cannot access memory at address 0x13c
> (gdb) thread 5 
> Thread ID 5 not known.
Thanks for the bt, I'll look at it.

> I'm having a bit of trouble getting more debug info.  It seems the usual
> trick of "./configure --enable-debug" doesn't work with this package.
Yep, that's an option I should add.

> Neither does overriding the CC environment variable.  I'll hack up the
I think a "-O3" or something like it must be in the makefile, and the
"make install" command strips the binary when you install it. Simply
build the game from the source dir and run src/liquidwar, and it won't
be stripped.

> Makefiles a bit and see if I can get more info.  I'd really like to play
> this game on my SMP desktop box, coz its incredibly addictive =)
You got an SMP box? Lucky guy!
However, you should know that LW will use only one CPU for all its
calculus, so it won't run faster on a SMP than on a "standard" single
CPU machine - provided that you run LW only 8-)

Thanks for the bug report, and have a nice day.

Christian.

-- 
Christian Mauduit

http://www.ufoot.org
mailto:address@hidden