Assigning to Henry to decide which fortunate engineer gets to work on this.
Having not used gsview on windows for a while (many years), I was surprised to see recently that it now offers three choices of driving printers under windows (it used to be just one...), and the default isn't "windows gdi" any more. Although the last version of gsview I used would have been 4.8 or 4.9, and current being 5.0, which AFAIK, aren't that different. windows gdi not being the default any more might be relevant to gsprint, too. (gsprint is part of gsview, not part of ghostscript).
The consensus is there really isn't anything we can about this with the information provided. If the customer can reproduce the problem with the gsprint command line and a postscript or pdf test file and a prescribed windows configuration (to include amount of RAM installed) we'll take a look. From the description it sounds like the customer can't reproduce the problem. Is that the case?
I've contacted the customer asking for additional information on how to reproduce the problem; I'll reopen the bug when I receive it.