On Mon, Sep 21, 2009 at 02:51:50AM -0500, Yaakov (Cygwin/X) wrote: >On 20/09/2009 21:24, Christopher Faylor wrote: >>This should be fixed in the next Cygwin snapshot and, subsequently, in >>the next release. >> >>If you could test this and confirm that the snapshot fixes the problem >>I'll roll a new release. It turns out to be a pretty serious bug. > >For me this fixes the problems reported here and those we discussed >earlier but were unable to resolve. I agree on the seriousness of >this, but what exactly helped you nail it down?
gdb said that the failure was coming from libxcb-1.dll so I rebuilt libxcb-1.dll with debugging information and with a version of libcygwin.a containing debugging symbols. Then I could see that it was dying in cygwin in a place that has seem a fair amount of churn recently. I could see that it had nothing to do with libxcb-1.dll. The fix for the problem wasn't immediately obvious and it is very hacky. I hope it can be removed in a couple of years when everything that was built using a buggy version of libcygwin.a is gone. The problem came from one of the X libraries that, if it had been relinked recently, would not have manifested a problem. I'm embarrassed to admit that I didn't record which library had the problem, though. Tracking it down in gdb proved to be sort of time-consuming and now the window with the information is gone. I can say that any DLL built with cygwin-1.7.0-51 - cygwin-1.7.0-56 is probably suspect. That's 2009-07-13 - 2009-08-13 . cgf -- 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
