If it is of any use, the versions I have installed, bash/sh 4.1.10(4), have the same length and differ in two byte positions, by one bit in each case.
These differences may just reflect the different name or a slightly different time at which the .exe was constructed as part of a build process. I suspect that bash is acting differently based on the name via which it was called and that the executable is really the same in both cases. On a true linux system they might be the same file, accessed via a hard link. The behavioral difference presumably could be traced down in the source code to bash. I have no idea how hard or subtle that may be. It might be a question that upstream maintainers could answer. I suppose it could still be some kind of very subtle interaction between bash/sh and cygwin, but you'd need to narrow down to particular library or system calls. I can't recall if you have tried using strace to trace system calls of the two invocations to see if you can find a difference there. I am not sure what you would look for or if it would even show up in a syscall trace ... Regards -- Eliot -- 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
