I am using Debian Lenny 5.0.3 with a stock 2.6.26-2-amd64 kernel, and have a ~/Documents/HTMLS directory of 273.2MB (21590 files, 1063 sub-folders) which according to konqueror shows 1130 items - 572 files (60.8MB Total) - 558 folders, these were created by 'Save Page As' in Iceweasel 3.5.11.

I am trying to copy the contents of this directory to a similar directory on a 2GB USB thumbdrive /dev/sdb1 which was partitioned and formatted as fat32 by Qparted.

Problem is that the copying ceases after about 6 files transfer. I found that (as I am sure you know) the named.html files come with a matching named_files folder, and often there are what appear to be invalid characters such as '*','?',and ':' in the filenames in said named_files folders.

After exhaustive googling [I know we all say that] I found an instruction:

that changes the name of the files, replacing said characters with '_', but whilst I can get pax from my repos, I really don't want to 'archive' the files - since I understand that pax was created to bridge a war between tar and cpio - because I want to be able to read the html files on an old (not connected to the Internet) WinXP tablet.

Sooo, I believe that I need to create a script, that scans all the filenames, greps and seds to replace said 'unacceptable' characters. I am assuming that Firefox on the Tablet PC will be able to open the htmls if I can get them onto the thumb drive.

Thanks in advance for any hints or tips that you can provide, As you have no doubt appreciated I ain't no programmer nor script kiddie!

Edit: Are there any other known characters in filenames that M$ file systems can't handle?

Not sure if it will work, but you could try formatting the USB drive to ext2/3 and use ext2fsd on Windows to read the drive. The question is will Windows accept the "special" characters on the drive.

Edit: I think Windows doesn't like / and \ either. And under some conditions everything that is not in the local codepage (problems possible under DOS and with applications using old MS API)