[Monotone-devel] Re: Rosters bug?

[Bruce Stephens]

Bruce Stephens <address@hidden> writes:

[...]

> I see the same on net.venge.monotone.  I run "make check" fairly
> regularly, and I'm fairly sure I didn't see it a couple of days ago,
> so I imagine something very recent in unstable has changed it.  Maybe
> the binutils change?

OK, if I build without -O2
(net.venge.monotone/oxygene/2005-11-21T10:53:42/i:4b73fd), I get a
segfault running unit_tests:

monotone: checked_globish_to_regex: input = 'a'
monotone: checked_globish_to_regex: output = 'a'
monotone: checked_globish_to_regex: input = 'b'
monotone: checked_globish_to_regex: output = 'b'
monotone: checked_globish_to_regex: input = 'c'
monotone: checked_globish_to_regex: output = 'c'
monotone: checked_globish_to_regex: input = '{a,b}?*\*|'
monotone: checked_globish_to_regex: output = '(a|b)..*\*\|'

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1223153984 (LWP 13948)]
0xb7dd9c01 in __gnu_cxx::__pool<true>::_M_reclaim_block ()
   from /usr/lib/libstdc++.so.6

#0  0xb7dd9c01 in __gnu_cxx::__pool<true>::_M_reclaim_block ()
   from /usr/lib/libstdc++.so.6
#1  0xb7f19875 in boost::basic_regex<char, boost::regex_traits<char, 
boost::cpp_regex_traits<char> > >::assign ()
   from /usr/lib/libboost_regex-gcc-mt-1_33.so.1.33.0
#2  0x082c799f in boost::basic_regex<char, boost::regex_traits<char, 
boost::cpp_regex_traits<char> > >::operator=<std::char_traits<char>, 
std::allocator<char> > (this=0xbf9d68b8, address@hidden) at basic_regex.hpp:274
#3  0x082c4429 in globish_matcher (this=0xbf9d68b8, address@hidden, 
    address@hidden) at globish.cc:128
#4  0x082c4fbb in globish_matcher_test () at globish.cc:200

[...]