So here is the latest from the cygwin folks: > -----Original Message----- > From: [email protected] [mailto:[email protected]] On > Behalf Of Eric Blake > Sent: Wednesday, March 17, 2010 2:41 PM > To: [email protected] > Subject: Re: 1.7.1: cvs version built in is unstable > > On 03/17/2010 03:39 PM, Roe, Kevin L. wrote: > > This appears to be related to another problem I am encountering. I > have described it in the thread "cp: skipping file 'file.txt', as it > was replaced while being copied" > > > > The same drive that has the CVS issue has the "cp" issue and the > other drive has neither issue. > > The problem is thus that the drive gives bogus inode values, but cygwin > does not know how to detect that those values are bogus to use its > fallback based on file name hashing. > > Is this with cygwin 1.7.1 or a snapshot? Corinna has done some work in > CVS for various buggy file systems, but you may have encountered yet > one more. > > -- > Eric Blake [email protected] +1-801-349-2682 > Libvirt virtualization library http://*libvirt.org
-----Original Message----- From: Larry Jones [mailto:[email protected]] Sent: Wednesday, March 17, 2010 11:31 AM To: Roe, Kevin L. Cc: [email protected] Subject: Re: "No such file" Roe, Kevin L. writes: > > S-> RCS_checkout (/usr/gapps/AEAG/cvs_modules/Utilities/p_change.pl,v, 1.8, , > , (function)) > cvs checkout: cannot write Utilities/p_change.pl: No such file or directory Hmmmm, that looks like some kind of a bug/incompatibility in cygwin. You need to ask the cygwin folks what would cause open(".new.p_change.pl", O_WRONLY | O_CREAT | O_TRUNC, 0777) to fail with errno set to ENOENT. -- Larry Jones I wonder if you can refuse to inherit the world. -- Calvin
