[Pachi] MAC regression bug: segmentation fault

Petr Baudis pasky at ucw.cz
Fri May 27 22:13:40 CEST 2011


  Hi!

On Fri, May 27, 2011 at 07:38:12PM +0100, jonathan chetwynd wrote:
> not sure if there is a regular bug filing app in place....

  The mailing list. ;-)

> in the meantime
> 
> 2 bugs on new 15" mac pro 4 core one processor
> latest OS X, Xcode and fully up to date.
> both pachi built fine with no errors using MAC=1 only edit:
> 
> 4df88f81de79620926a27410e0948fa4092c5bca
> 
> ./pachi
> runs fine, 
> 
> and
> 
> ./pachi threads=2
> runs fine
> 
> but 
> 
> ./pachi threads=4 causes segmentation fault
> on genmove black  
> ie 9x9 standard

Hmm, curious. Would it be possible to run Pachi in a debugger, or
analyze the core dump? At least to get a backtrace (i.e. where within
the code Pachi crashed). I'm not familiar with Xcode, but if there is
'gdb' tool installed, just

	gdb --args ./pachi threads=4
	> run
	... let it crash ...
	> bt

> regression
> 
> 75dd178099bef5f3d426040d124bbd56a2b2aafd
> 
> ./pachi causes segmentation fault
> on genmove black  
> ie 9x9 standard

Most curious! Is 'genmove black' the very first command, or do you issue
some initialization commands as well?

The only relevant change is addition of 'fixed_handicap' to the list of
known commands, but I have no idea how this could be related.

-- 
				Petr "Pasky" Baudis
UNIX is user friendly, it's just picky about who its friends are.


More information about the Pachi mailing list