On 2014-01-23 04:32, Corinna Vinschen wrote:
On Jan 22 23:20, Yaakov (Cygwin/X) wrote:
Now the strange thing about this module is that it has a constructor
which calls pthread_once() on a function which creates a recursive
mutex, so just dlopen()ing is enough. Unfortunately, my attempts to
reproduce this with an STC have not been successful.
Corinna, cgf, any insights?
Except for the content of object apparently being off, no. It would
be helpful to build p11-kit with debugging and without optimization
and then step right into it, otherwise it's pretty hard to say
anything useful.
And, of course, without optimization it works. So, I managed to narrow
this down to the ctor itself: compiling it with -O0 or -O1 works, but
not with -O2 or -O3; other functions seem to be fine at -O2. The
diffstat between the resulting assemblies of that file is huge, so I
haven't diagnosed the actual cause yet. Feel free to let me know if
you're interested. :-)
In the meantime, the workaround is easy, so I just uploaded
p11-kit-0.18.7-2 and ca-certificates-1.96-2 for x64 only. If you are
running 64-bit Cygwin on Windows 8.0 x64, those releases should fix this
for you.
Yaakov
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
