Re: [Monotone-devel] sketch of i18n specification

[Kevin Smith]

graydon hoare wrote:
> Ori Berger <address@hidden> writes:
>
> > On Windows, situation's much worse -- you can't have any device name
> > as a file name (e.g., "nul", the windows equivalent to "/dev/null",
> > virtually exists in every directory on every device) --
> >
> > but that can't be solved by placing structural limits on file names.
>
>
> hm. this is a curious situation! can you give more details about funny
> things in windows filesystems? 

I can't tell you a lot about modern stuff, but back in the DOS days, 
there were several devices that you could refer to as if they existed in 
every directory. The ones I remember are: nul, prn, con, lpt1, lpt2, 
lpt3, com1, com2, com3...

...yahoo search in progress... Here's one website for you:
  http://www.tcs.org/ioport/jun98/driver2.htm

Here is perhaps a better historical explanation:
<http://blogs.gotdotnet.com/raymondc/commentview.aspx/1f7e4f66-2889-4238-8962-37899528e211>

This message is also nice and scary:
  http://mail.gnu.org/archive/html/bug-gnu-utils/2001-07/msg00044.html

"In addition, the file-name extension is ignored when the basename
matches.  So `aux.lst', `prn.c', `con.foo', and an infinite number of
other similar names--all of them are prone to this problem.  Some of the
devices will actually wedge the DOS box if you try to extract a file by
that name; kids, don't try that at home!"

Oy.

The good news is that I use a Windows box at work, and have not had 
problems with any of these for at least a decade. Apparently I've been 
lucky.

Kevin