I'm guessing there's some process (apparently running on a Windows box) that "autogenerates" this flat file. If your perl script is going to run on that same box, you might save yourself some trouble by leaving out the initial separate process, and just using the perl script to scan the path(s) of interest, and collect the file name, date and size information for each.

Depending on the "big picture" (whatever it is you are really trying to accomplish, how many directories there are, etc), you could either use opendir and readir, or some variant of File::Find.

One of those things, combined with stat (or the -X functions) and maybe POSIX should make for a fairly stable and portable solution.