On Thu, Aug 06, 2009 at 12:45:31PM -0400, Christopher Faylor wrote: >On Thu, Aug 06, 2009 at 06:37:34PM +0200, Corinna Vinschen wrote: >>On Aug 4 18:36, Christopher Faylor wrote: >>> On Tue, Aug 04, 2009 at 02:00:42PM -0400, Christopher Faylor wrote: >>> >On Tue, Aug 04, 2009 at 04:30:49PM +0000, Eric Blake wrote: >>> >>> I'm still seeing some weird stuff in the latest snapshot. This is >>> >>> a fresh install, with nothing other than the Base category + git, >>> >>> and substituting just cygwin1.dll. >>> >>> >>> >>> $ git clone git://git.sv.gnu.org/libsigsegv.git >>> >>> Initialized empty Git repository in /home/eblake/libsigsegv/.git/ >>> >>> fatal: write error (Socket operation on non-socket) >>> >> >>> >>I'm seeing this behavior even as far back as snapshot 20090706, so it >>> >>may be some BLODA on this particular machine, rather than a new >>> >>failure in the latest snapshot. Other than this weird behavior of >>> >>git, I haven't seen anything else misbehave with the 20090804 snapshot. >>> > >>> >Does that mean you don't see the problem with the 20090704 snapshot or >>> >that you haven't tested that far back? >>> > >>> >I'll try this command out as soon as I can. >>> >>> I tried it and it failed as you indicated. >>> >>> The next snapshot should fix the problem. >> >>FWIW, if I change the send/recv buffer sizes back to 65536 in fdsock, >>it still works for me, every time I try it. I tried on XP SP3, >>Server 2008, and Windows 7. >> >>Is it possible that this is actually a BLODA problem? > >I was, until yesterday, using the McAffee virus checker. I'll try it again >with that turned off.
If I set the buffer size up to 65536 it still fails. If I set it down it still succeeds. FWIW, my previous tests were with Windows XP SP2. Given that Eric is still reporting problems, I guess it could still be BLODA though. I can send cygcheck output if that helps. 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
