From: "Mauricio Fernández" Date: 2003-03-27T18:23:09+09:00 Subject: Re: Debugging Ruby itself. On Thu, Mar 27, 2003 at 02:29:58PM +0900, John Carter wrote: > For various reasons I'm using the CVS latest version of Ruby.... > > Now I have 3500 lines of _very_ _very_ threaded (about 100 threads) ruby > doing lots and lots of socket I/O to strange exotic hardware. > > Ruby itself sometimes crashes (not in the same place). > > So I haven't been able to produce a small test set. I can't send Matz all > the weird hardware, so I'm on my own here.... Does the bug look like the phantom one I found? http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/66166 Something like ./dict.rb:316: [BUG] gc_sweep(): unknown data type 28 > Any Ruby specific suggestions for hunting this type of bug? > > I'm on RedHat Linux 8.0 on a Pentium 4 using gcc 3.2.2 It could be GCC's fault. I was doing something heavily threaded too (from 20 to 150+ threads) which opened lots of connections to a dictd server and found that the problem would only happen when compiling with gcc 3.2, and moreover only when targeting i386, but not sure on this one, have forgotten :-) -- _ _ | |__ __ _| |_ ___ _ __ ___ __ _ _ __ | '_ \ / _` | __/ __| '_ ` _ \ / _` | '_ \ | |_) | (_| | |_\__ \ | | | | | (_| | | | | |_.__/ \__,_|\__|___/_| |_| |_|\__,_|_| |_| Running Debian GNU/Linux Sid (unstable) batsman dot geo at yahoo dot com * Jes wonders why so many people in here uses fooZZZZZ and foo_sleeping nicks Jes: Because they are sleeping? -- Seen on #Linux