identical scripts don't behave identically

[roc97007 <ronc at europa dot com>]

We have been seeing bizarre results from Cygwin 1.5.25-15 on Windows Server
2003 R2.  I have reduced the problem to the following test:

I have two files with identical permissions located in the same directory:

-rwxrwxrwx 1 christian.ron mkpasswd 16 Jul  8 14:56 aab.sh
-rwxrwxrwx 1 christian.ron mkpasswd 18 Jul  8 14:38 aaa.sh

They have identical contents:

$ cat aaa.sh
#!/bin/bash
ls

$ cat aab.sh
#!/bin/bash
ls

They have identical ACLs

$ getfacl aaa.sh
# file: aaa.sh
# owner: christian.ron
# group: mkpasswd
user::---
group::---
mask:rwx
other:rwx

$ getfacl aab.sh
# file: aab.sh
# owner: christian.ron
# group: mkpasswd
user::---
group::---
mask:rwx
other:rwx

Now, nothing up my sleeve, when aaa.sh is executed:

$ ./aaa.sh
: No such file or directory

When aab.sh is executed:

$ ./aab.sh
Copy of ad_synch.sql   blarg.sh       foo.sh                   xxx.sh
Copy of enchilada.sh~  broken.sh      foob1.sh                 xxx1.sh
aaa.sh                 enchilada.bat  foob2.sh
aab.sh                 enchilada.sh~  test.sql
ad_synch.sql           enchilada.sql  the_whole_enchilada.txt

To summarize, as far as I can tell the files are identical, but one works
and the other can't seem to find the current directory.  If, in the broken
script, I substitute any command that needs to read or write to or from the
file system, it will always fail with "file not found" or "no such file or
directory".  The other script with the exact same commands (assuming they
are legal and reasonable) will always work.

Obviously there is a difference between the files.  I can't figure out what
it is.  The issue is demonstrably a characteristic of the file, because if I
delete or rename aaa.sh and then "cp aab.sh aaa.sh", then they will both
work.

This is driving me crazy.  We are under a deadline, and some scripts work
and some do not, seemingly randomly.  Our current work-around is to copy a
script that works to a different name and then replace the content.  This is
a ridiculous workaround, and since we don't know what the mechanism is we
have no idea if it'll still work tomorrow or next week.

Please help!
-- 
View this message in context: http://www.nabble.com/identical-scripts-don%27t-behave-identically-tp24400289p24400289.html
Sent from the Cygwin list mailing list archive at Nabble.com.


--
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