On Thu, Jun 06, 2002 at 11:50:56PM -0500, xOr wrote:
> On Thu, Jun 06, 2002 at 11:28:10PM -0500, xOr wrote:
> > Found a way to make blackbox block in its select() call.
> > 
> > Theres a comment above it say this is possible but near impossible..
> > seems not to be the case :|  It does it perfecetly consistantly on my
> > machine if I follow the right steps.
> > 
> > This is, of course, with the latest cvs version. But should work for the
> > last few (all?) alphas too.
> > Also, I am using the cvs version of bbkeys. With bbkeys 0.8.4, this
> > doesn't happen for me. However, it seems that blackbox blocking is still
> > a problem regardless. I will attempt to track down what change in bbkeys
> > caused it.
> 
> OK, I figured it out, you have to compile bbkeys with --enable-debug it
> seems. Very strange.. 0.8.4 does it then also.

So, further narrowing has discovered that it is the -DDEBUG flag that
makes bbkeys make blackbox block. However, if all the #ifdef DEBUG's are
removed from the code and -DDEBUG is not used, it doesnt make blackbox
block. So something in a system header when bbkeys is compiled with
-DDEBUG is causing blackbox to block with the below method on my machine
it seems. Freaky..

> > heres the .xinitrc:
> > 
> > bbkeys -i&
> > gkrellm -w &
> > exec blackbox
> > 
> > Then, once blackbox has loaded:
> > launch a terminal.
> > In the terminal run:
> > % bbkeys
> > The X server should be entirely frozen at this point.
> > 
> > gdb reports this backtrace.
> > 
> > 0x402547ce in select () from /lib/libc.so.6
> > (gdb) bt
> > #0  0x402547ce in select () from /lib/libc.so.6
> > #1  0xbffff4a4 in ?? ()
> > #2  0x0809f5af in main (argc=1, argv=0xbffff864) at main.cc:160
> > 
> > That's with debug symbols enabled.. :\ not sure why it gets a ??.
> > 
> > Can anyone else reproduce this?
> > 
> 
> xOr
> -- 
> I am damn unsatisfied to be killed in this way.




-- 
I am damn unsatisfied to be killed in this way.

Attachment: msg07356/pgp00000.pgp
Description: PGP signature

Reply via email to