Bus Error multi-handle teardown? about curb HOT 5 CLOSED

I also get an error when running the unit tests that seems to be related:

  1) Error:
test_with_success_cb_with_404(TestCurbCurlMulti):
TypeError: wrong argument type Fixnum (expected Data)
    /Users/ian/development/atreia/deps/curb/tests/tc_curl_multi.rb:6:in `garbage_collect
    /Users/ian/development/atreia/deps/curb/tests/tc_curl_multi.rb:6:in `teardown'

This isn't consistently reproducible, unfortunately. It seems to occur in ~50% of the runs.

Oddly enough, if you add a puts statement to the end of the test, the error doesn't seem to occur as often, if at all. Race condition of some sort on cleaning up the multi handle?

from curb.

Here's some additional information, when running valgrind.

valgrind /home/taf2/.local/bin/ruby tests/bug_multi_segfault.rb 
==13822== Memcheck, a memory error detector.
==13822== Copyright (C) 2002-2008, and GNU GPL'd, by Julian Seward et al.
==13822== Using LibVEX rev 1884, a library for dynamic binary translation.
==13822== Copyright (C) 2004-2008, and GNU GPL'd, by OpenWorks LLP.
==13822== Using valgrind-3.4.1, a dynamic binary instrumentation framework.
==13822== Copyright (C) 2000-2008, and GNU GPL'd, by Julian Seward et al.
==13822== For more details, rerun with: -v
==13822== 
==13822== Invalid read of size 8
==13822==    at 0xBBBC95A: curl_multi_free (curb_multi.c:61)
==13822==    by 0x4E931AD: run_final (gc.c:2349)
==13822==    by 0x4E91532: finalize_list (gc.c:1661)
==13822==    by 0x4E936E6: rb_gc_call_finalizer_at_exit (gc.c:2464)
==13822==    by 0x4E7FD86: ruby_finalize_1 (eval.c:117)
==13822==    by 0x4E7FF19: ruby_cleanup (eval.c:157)
==13822==    by 0x4E80185: ruby_run_node (eval.c:234)
==13822==    by 0x40090A: main (main.c:35)
==13822==  Address 0x10 is not stack'd, malloc'd or (recently) free'd
==13822== 
==13822== Invalid read of size 2
==13822==    at 0x4F8A0D1: rb_vm_get_sourceline (vm.c:680)
==13822==    by 0x4F8A471: rb_sourceline (vm.c:769)
==13822==    by 0x4E7B9AF: rb_bug (error.c:230)
==13822==    by 0x4F1E25D: sigsegv (signal.c:605)
==13822==    by 0x523EE8F: (within /lib64/libpthread-2.10.1.so)
==13822==    by 0xBBBC959: curl_multi_free (curb_multi.c:61)
==13822==    by 0x4E931AD: run_final (gc.c:2349)
==13822==    by 0x4E91532: finalize_list (gc.c:1661)
==13822==    by 0x4E936E6: rb_gc_call_finalizer_at_exit (gc.c:2464)
==13822==    by 0x4E7FD86: ruby_finalize_1 (eval.c:117)
==13822==    by 0x4E7FF19: ruby_cleanup (eval.c:157)
==13822==    by 0x4E80185: ruby_run_node (eval.c:234)
==13822==  Address 0x64f489c is 4 bytes before a block of size 12 alloc'd
==13822==    at 0x4C2563E: malloc (vg_replace_malloc.c:207)
==13822==    by 0x4E8F811: vm_xmalloc (gc.c:624)
==13822==    by 0x4E8F9C1: ruby_xmalloc2 (gc.c:711)
==13822==    by 0x4F64C75: iseq_set_sequence (compile.c:1311)
==13822==    by 0x4F640FE: iseq_setup (compile.c:954)
==13822==    by 0x4F637AA: rb_iseq_compile_node (compile.c:503)
==13822==    by 0x4F75972: rb_iseq_new_with_bopt_and_opt (iseq.c:336)
==13822==    by 0x4F759D4: rb_iseq_new_with_opt (iseq.c:346)
==13822==    by 0x4F75839: rb_iseq_new (iseq.c:304)
==13822==    by 0x4F8CC64: Init_VM (vm.c:1911)
==13822==    by 0x4E9A4B0: rb_call_inits (inits.c:55)
==13822==    by 0x4E7FBC2: ruby_init (eval.c:66)
: [BUG] Segmentation fault
ruby 1.9.1p129 (2009-05-12 revision 23412) [x86_64-linux]

-- control frame ----------
==13822== 
==13822== Conditional jump or move depends on uninitialised value(s)
==13822==    at 0x4F8D02D: control_frame_dump (vm_dump.c:33)
==13822==    by 0x4F8D70F: rb_vmdebug_stack_dump_raw (vm_dump.c:176)
==13822==    by 0x4F8E12C: rb_vm_bugreport (vm_dump.c:575)
==13822==    by 0x4E7B8C1: report_bug (error.c:215)
==13822==    by 0x4E7B9CE: rb_bug (error.c:230)
==13822==    by 0x4F1E25D: sigsegv (signal.c:605)
==13822==    by 0x523EE8F: (within /lib64/libpthread-2.10.1.so)
==13822==    by 0xBBBC959: curl_multi_free (curb_multi.c:61)
==13822==    by 0x4E931AD: run_final (gc.c:2349)
==13822==    by 0x4E91532: finalize_list (gc.c:1661)
==13822==    by 0x4E936E6: rb_gc_call_finalizer_at_exit (gc.c:2464)
==13822==    by 0x4E7FD86: ruby_finalize_1 (eval.c:117)
==13822== 
==13822== Invalid read of size 2
==13822==    at 0x4F8A0D1: rb_vm_get_sourceline (vm.c:680)
==13822==    by 0x4F8D358: control_frame_dump (vm_dump.c:109)
==13822==    by 0x4F8D70F: rb_vmdebug_stack_dump_raw (vm_dump.c:176)
==13822==    by 0x4F8E12C: rb_vm_bugreport (vm_dump.c:575)
==13822==    by 0x4E7B8C1: report_bug (error.c:215)
==13822==    by 0x4E7B9CE: rb_bug (error.c:230)
==13822==    by 0x4F1E25D: sigsegv (signal.c:605)
==13822==    by 0x523EE8F: (within /lib64/libpthread-2.10.1.so)
==13822==    by 0xBBBC959: curl_multi_free (curb_multi.c:61)
==13822==    by 0x4E931AD: run_final (gc.c:2349)
==13822==    by 0x4E91532: finalize_list (gc.c:1661)
==13822==    by 0x4E936E6: rb_gc_call_finalizer_at_exit (gc.c:2464)
==13822==  Address 0x64f489c is 4 bytes before a block of size 12 alloc'd
==13822==    at 0x4C2563E: malloc (vg_replace_malloc.c:207)
==13822==    by 0x4E8F811: vm_xmalloc (gc.c:624)
==13822==    by 0x4E8F9C1: ruby_xmalloc2 (gc.c:711)
==13822==    by 0x4F64C75: iseq_set_sequence (compile.c:1311)
==13822==    by 0x4F640FE: iseq_setup (compile.c:954)
==13822==    by 0x4F637AA: rb_iseq_compile_node (compile.c:503)
==13822==    by 0x4F75972: rb_iseq_new_with_bopt_and_opt (iseq.c:336)
==13822==    by 0x4F759D4: rb_iseq_new_with_opt (iseq.c:346)
==13822==    by 0x4F75839: rb_iseq_new (iseq.c:304)
==13822==    by 0x4F8CC64: Init_VM (vm.c:1911)
==13822==    by 0x4E9A4B0: rb_call_inits (inits.c:55)
==13822==    by 0x4E7FBC2: ruby_init (eval.c:66)
c:0001 p:0000 s:0002 b:0002 l:002628 d:002628 TOP   
---------------------------
-- Ruby level backtrace information-----------------------------------------

-- C level backtrace information -------------------------------------------
0x4f8e26a /home/taf2/.local/lib/libruby.so.1.9(rb_vm_bugreport+0x179) [0x4f8e26a]
0x4e7b8c2 /home/taf2/.local/lib/libruby.so.1.9 [0x4e7b8c2]
0x4e7b9cf /home/taf2/.local/lib/libruby.so.1.9(rb_bug+0xf1) [0x4e7b9cf]
0x4f1e25e /home/taf2/.local/lib/libruby.so.1.9 [0x4f1e25e]
0x523ee90 /lib64/libpthread.so.0 [0x523ee90]
0xbbbc95a /home/taf2/work/curb/ext/curb_core.so [0xbbbc95a]
0x4e931ae /home/taf2/.local/lib/libruby.so.1.9 [0x4e931ae]
0x4e91533 /home/taf2/.local/lib/libruby.so.1.9 [0x4e91533]
0x4e936e7 /home/taf2/.local/lib/libruby.so.1.9(rb_gc_call_finalizer_at_exit+0x278) [0x4e936e7]
0x4e7fd87 /home/taf2/.local/lib/libruby.so.1.9 [0x4e7fd87]
0x4e7ff1a /home/taf2/.local/lib/libruby.so.1.9(ruby_cleanup+0x181) [0x4e7ff1a]
0x4e80186 /home/taf2/.local/lib/libruby.so.1.9(ruby_run_node+0x73) [0x4e80186]
0x40090b /home/taf2/.local/bin/ruby(main+0x4f) [0x40090b]
0x5d31a2d /lib64/libc.so.6(__libc_start_main+0xfd) [0x5d31a2d]
0x4007f9 /home/taf2/.local/bin/ruby [0x4007f9]

[NOTE]
You may encounter a bug of Ruby interpreter. Bug reports are welcome.
For details: http://www.ruby-lang.org/bugreport.html

==13822== 
==13822== ERROR SUMMARY: 4 errors from 4 contexts (suppressed: 6 from 2)
==13822== malloc/free: in use at exit: 2,949,375 bytes in 15,296 blocks.
==13822== malloc/free: 16,587 allocs, 1,291 frees, 3,507,212 bytes allocated.
==13822== For counts of detected errors, rerun with: -v
==13822== Use --track-origins=yes to see where uninitialised values come from
==13822== searching for pointers to 15,296 not-freed blocks.
==13822== checked 2,165,304 bytes.
==13822== 
==13822== LEAK SUMMARY:
==13822==    definitely lost: 761 bytes in 14 blocks.
==13822==      possibly lost: 51,928 bytes in 925 blocks.
==13822==    still reachable: 2,896,686 bytes in 14,357 blocks.
==13822==         suppressed: 0 bytes in 0 blocks.
==13822== Rerun with --leak-check=full to see details of leaked memory.
Aborted

from curb.

Okay, I have a fix for the segv, if you pull and test on Mac OSX with ruby 1.9 that'd be very helpful thanks you!

from curb.

Looks good on OS X! Thanks for the quick fix :)

from curb.

Great, I'll spin a new release in a few

from curb.