A very poor search made me lost an entire afternoon conducting regression tests when the error and the "bad" commit were already identified... my fault entirely [Evil or Very Mad]. Well, at least i learned how to make a regression test. For those who are affected by this particular bug, you can follow further developments in http://bugs.winehq.org/show_bug.cgi?id=23902 . The patch published there by the original author of the 'bad' commit apparently solved my crashes.
