Ruby » Backport87

i've now compiled ruby 1.8.8dev (2010-03-13) (i.e. ruby_1_8 branch) and it's crashing with the same backtrace.

This problem might be related to #2603 and you had better try patches attached to #2739.
I received a private mail, but I write to here for memorandum record.

Taca: thanks, i've confirmed the problem is still present in ruby_1_8 (r26956), there should be the patches from #2739 already included

Yui: i've installed the 'ruby-mysql' gem instead of the 'mysql' gem now (and also made an upgrade of mysql from 5.0.90 to 5.1.44, but this probably is not significant). i can still bring the app to crash and it seems to me, now it is pure ruby (without mysql code involved):

Program received signal SIGSEGV, Segmentation fault.
0xb98b8c27 in pthread_exit (retval=0x1) at /usr/src/lib/libpthread/pthread.c:526
526                     }
(gdb) bt
#0  0xb98b8c27 in pthread_exit (retval=0x1) at /usr/src/lib/libpthread/pthread.c:526
#1  0xb98b8d89 in pthread__cancelled () at /usr/src/lib/libpthread/pthread.c:896
#2  0xb98b4b39 in fcntl (fd=11, cmd=3) at /usr/src/lib/libpthread/pthread_cancelstub.c:185
#3  0xbbb5078f in io_fwrite (str=3085207420, fptr=0xb9643da0) at io.c:454
#4  0xbbb57c7f in io_write (io=3085802960, str=3085207420) at io.c:602
#5  0xbbb2a82e in call_cfunc (func=0xbbb57b70 <io_write>, recv=3085802960, len=893832, argc=-1145210616, argv=0xbfbfb4c0) at eval.c:5861
#6  0xbbb35109 in rb_call0 (klass=3145897400, recv=3085802960, id=7385, oid=7385, argc=-1145210616, argv=0xbfbfb4c0, body=0xbb828f00, flags=<value optimized out>) at eval.c:6005
#7  0xbbb35f4f in rb_call (klass=3145897400, recv=3085802960, mid=7385, argc=1, argv=0xbfbfb4c0, scope=0, self=3085801380) at eval.c:6257
#8  0xbbb32165 in rb_eval (self=3085801380, n=<value optimized out>) at eval.c:3568
#9  0xbbb333f6 in rb_eval (self=3085801380, n=<value optimized out>) at eval.c:3382
#10 0xbbb35aee in rb_call0 (klass=3138364640, recv=3085801380, id=7385, oid=7385, argc=0, argv=0xbfbfba70, body=0xbb0faf1c, flags=<value optimized out>) at eval.c:6160
#11 0xbbb35f4f in rb_call (klass=3138364640, recv=3085801380, mid=7385, argc=1, argv=0xbfbfba70, scope=0, self=3103098700) at eval.c:6257
#12 0xbbb32165 in rb_eval (self=3103098700, n=<value optimized out>) at eval.c:3568
#13 0xbbb3eaa4 in rb_yield_0 (val=3085207420, self=3103098700, klass=<value optimized out>, flags=<value optimized out>, avalue=0) at eval.c:5167
#14 0xbbb4014d in rb_yield (val=3085207420) at eval.c:5251
#15 0xbbb9fea5 in rb_str_each_line (argc=0, argv=0x0, str=3083915440) at string.c:3900
#16 0xbbb2a852 in call_cfunc (func=0xbbb9fd30 <rb_str_each_line>, recv=3083915440, len=893832, argc=-1145210616, argv=0x0) at eval.c:5855
#17 0xbbb35109 in rb_call0 (klass=3145929720, recv=3083915440, id=4961, oid=4961, argc=-1077951308, argv=0x0, body=0xbb830660, flags=<value optimized out>) at eval.c:6005
#18 0xbbb35f4f in rb_call (klass=3145929720, recv=3083915440, mid=4961, argc=0, argv=0x0, scope=0, self=3086564500) at eval.c:6257
#19 0xbbb32165 in rb_eval (self=3086564500, n=<value optimized out>) at eval.c:3568
#20 0xbbb3924a in rb_block_pass (func=0xbbb341c0 <call_block>, arg=3217015892, proc=3085212040) at eval.c:9288
#21 0xbbb322fe in rb_eval (self=3086564500, n=<value optimized out>) at eval.c:9362
#22 0xbbb35aee in rb_call0 (klass=3105092260, recv=3086564500, id=4001, oid=4001, argc=0, argv=0x0, body=0xb915052c, flags=<value optimized out>) at eval.c:6160
#23 0xbbb35f4f in rb_call (klass=3105092260, recv=3086564500, mid=4001, argc=0, argv=0x0, scope=0, self=3103098700) at eval.c:6257
#24 0xbbb32165 in rb_eval (self=3103098700, n=<value optimized out>) at eval.c:3568
#25 0xbbb33d67 in rb_eval (self=3103098700, n=<value optimized out>) at eval.c:3296
#26 0xbbb3396b in rb_eval (self=3103098700, n=<value optimized out>) at eval.c:3432
#27 0xbbb35aee in rb_call0 (klass=3138089540, recv=3103098700, id=26609, oid=26609, argc=0, argv=0xbfbfceb0, body=0xbb001ed0, flags=<value optimized out>) at eval.c:6160
#28 0xbbb35f4f in rb_call (klass=3138089540, recv=3103098700, mid=26609, argc=2, argv=0xbfbfceb0, scope=0, self=3103098900) at eval.c:6257
#29 0xbbb32165 in rb_eval (self=3103098900, n=<value optimized out>) at eval.c:3568
#30 0xbbb3eaa4 in rb_yield_0 (val=3103098700, self=3103098900, klass=<value optimized out>, flags=<value optimized out>, avalue=0) at eval.c:5167

One quesition.

How did you install ruby on NetBSD?  Have you installed via pkgsrc?
If so, do you use latest pkgsrc (pkgsrc-current) or branch (pkgsrc-2009Q1)?
And what about MySQL?

If you install MySQL via pkgsrc, please use mysql5-client/mysql5-server instead of
mysql51-client/mysql51-server.  mysql51 in pkgsrc might have some problem and
mysql5 is more stable now.

I need switching between different versions of ruby and combinations of gems, so i use hand-compiled ruby. I compile ruby from the 1.8 branches with the following (with appropriate target_dir):

target_dir="/opt/ruby-1.8-20100316-rails_2.3"
env CFLAGS="-g -march=i686 -O2 -pipe -fomit-frame-pointer" ./configure --prefix=${target_dir} --mandir=${target_dir}/man --datadir=${target_dir}/share --enable-shared --enable-static
make 
make install

(plus the ruby_1_8 snapshot needed 'autoconf configure.in > configure')

The first backtrace was with mysql5 (client, server) from pkgsrc from branch pkgsrc-2009Q4. I really don't see, how relevant is the mysql code to the second backtrace, as i wrote in previous: i am using the 'ruby-mysql' gem instead of the 'mysql' gem now (yes, it was mysql51 (client, server) from current pkgsrc branch (other packages: pkgsrc-2009Q4)).

here is another backtrace (still the ruby_1_8 snapshot from yesterday):

Program received signal SIGSEGV, Segmentation fault.
0xb98b8c27 in pthread_exit (retval=0x1)
    at /usr/src/lib/libpthread/pthread.c:526
526                     }
(gdb) bt
#0  0xb98b8c27 in pthread_exit (retval=0x1) at /usr/src/lib/libpthread/pthread.c:526
#1  0xb98b8d89 in pthread__cancelled () at /usr/src/lib/libpthread/pthread.c:896
#2  0xb98b3de3 in accept (s=10, addr=0xbfbfd780, addrlen=0xbfbfd800) at /usr/src/lib/libpthread/pthread_cancelstub.c:127
#3  0xbb9cd28e in s_accept (klass=3141521780, fd=10, sockaddr=0xbfbfd780, len=0xbfbfd800) at socket.c:1517
#4  0xbb9cd6c7 in tcp_accept (sock=3103098860) at socket.c:1551
#5  0xbbb2a83a in call_cfunc (func=0xbb9cd660 <tcp_accept>, recv=3103098860, len=893832, argc=-1145210616, argv=0x0) at eval.c:5858
#6  0xbbb35109 in rb_call0 (klass=3141521660, recv=3103098860, id=17857, oid=17857, argc=1, argv=0x0, body=0xbb3fd0d4, flags=<value optimized out>) at eval.c:6005
#7  0xbbb35f4f in rb_call (klass=3141521660, recv=3103098860, mid=17857, argc=0, argv=0x0, scope=0, self=3103098900) at eval.c:6257
#8  0xbbb32165 in rb_eval (self=3103098900, n=<value optimized out>) at eval.c:3568
#9  0xbbb30de1 in rb_eval (self=3103098900, n=<value optimized out>) at eval.c:3772
#10 0xbbb333f6 in rb_eval (self=3103098900, n=<value optimized out>) at eval.c:3382

try to run configure with --enable-pthread

that seems to have helped, thanks!

so far no problems with:
env CFLAGS="-g -march=i686 -O2 -pipe -fomit-frame-pointer" ./configure --prefix=${target_dir} --mandir=${target_dir}/man --datadir=${target_dir}/share --enable-pthread --enable-shared --with-tcllib=none --with-tklib=none