Reliable old script loses data on new Cygwin installation

Thomas Baker
Wed Nov 14 09:07:00 GMT 2007

On Tue, Nov 13, 2007 at 05:59:01PM +0100, Thomas Baker wrote:
> > The speed is not the problem, it could be the usual suspect: an
> > anti-virus, unlikely because the data written is not executable but it
> > could be adding an extraneous delay between data written and data read.
> I'll ask the guy who installed the operating systems whether this
> looks possible.

These machines have an anti-virus program, but the same one I have 
been using for the past two or three years.  The filenames either have no 
extension, or ".txt".

> > The only solid evidence is the error message from Windows, and it says
> > "device", that means that the hard disk is having problems (it could be
> > the driver) but not your script or any program.  Did you do anything
> > special to the hard disks on all 3 machines? something like run an
> > optimizer or tune parameters?
> I'll find out.

The guy who set up the machines did not fiddle with the hard
disk at all.  Two of the machines were brand-new.  On these,
he did not touch the hardware - only swapped the installed
German-language version of XP for my own version of XP
in English.  Then I installed Cygwin from the net.

I'm a bit at a loss as to what to do next.  I use this one 
script so extensively (50 times per day?) that I would sooner
move to a Linux or Macintosh system than continue on Windows
without Cygwin - my whole working style depends on it.

I work on a laptop with Cygwin and an up-to-date version of
XP, and thankfully the script continues to run fine there.
Having (seemingly) eliminated SATA/ATA, fast/slow, and
old-XP/up-to-date-XMP as possible explanations for the errors,
the only really obvious remaining difference between my laptop
(where the script works) and the four desktop computers
(where the script does not work reliably) is that the four
desktop computers have the latest version of Cygwin and the
laptop has a version that I have not updated in awhile (and
now I do not dare!).

I would be grateful if someone could suggest a way to test
this.  Would it make sense (and is it possible) to replace
the Cygwin kernel or the "coreutils" package (because of
"mv") with earlier versions and see if the script works then?
The version of "pdksh" is the same between the laptop and the
desktops, so I assume this can be eliminated as the cause?
The script itself is not long, and the data sets are also
just a few pages long, so I could prepare a tar file with a
simple test installation if anyone would like to try it on
their machines.


Tom Baker - -

Unsubscribe info:
Problem reports:

More information about the Cygwin mailing list