Actually, in this case, the file size is 6089617325 bytes, but it reports 1794650029 bytes.
Created attachment 1757 [details]
Ethereal capture of Cygwin ls -l largefile
Problem also visable in explorer, so not Cygwin specific.
Note in the packet capture that the return from query file standard info shows:
Allocation Size: 6096420864
End of File: 179650029
Created attachment 1758 [details]
Same Ethereal snnop of Cygwin ls -l a few minutes later
Now, query file standard info shows:
Allocation size: 6096420864
End of File: 6089617325
the correct information.
Created attachment 1759 [details]
1 minute of log level 10
This should be 1 minute of log level 10 output during the problem of Cygwin bash doing:
while :; do ls -l /cygdrive/z/release/env9/commercial/qa_wa/uk.txf; done
Sorry, but I can't seem to correlate this log to the Ethereal snoop so as to identify the cycle and pare down the log any further.
Note that the log and the Ethereal do not exactly match. The process was the same, but the time was different.
Let me know if I can provide anthing additional, as this is readily reproducable. Thanks.
I have this exact same problem on my nslu2 (http://www.nslu2-linux.org/wiki/Main/HomePage) with samba 3.0.20-r4 installed. Is any work being done on this bug?