Here is a resubmit of my earlier patch. Please disregard that one. In this patch, I: - removed the superfluous code in resolveFile() - made both resolveFile and normalize key off of a ":" anywhere in the string instead of ":" in the second position only in determining if they have an absolute path name. - fixed normalize to comply to the spec, handling both Windows drives and NetWare volumes in the same code.
I am still wondering if the way that I coded this is indeed the intended behavior. Now, both normalize and resolveFile would handle drive names of "SYS:/" on a windows machine. But, as somebody mentioned, it could be that somebody is using emulation software and so those paths would eventually be correct. If not, the code would fail otherwise than in here. These routines will just standardize the name (in the case of normalize and in the case of resolveFile when the filename parameter is actually just a drive letter), and/or actually resolve the file in other cases, and then pass that name on to the calling routine, where the given file may not actually exist. An expanded FileUtilsTest will follow. Jeff Tulley ([EMAIL PROTECTED]) (801)861-5322 Novell, Inc., the leading provider of Net services software.
FileUtilsforNW.patch
Description: Binary data
