This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: 3 bugs


On my bug report
> My program could create a file "xy." (with a final dot) in the cygwin 
> environment using a link() call, but I could not manage by any means 
> to use, remove, or rename that file afterwards.
there were several replies.

: I encounter a similar problem.  Apparently, files with a trailing "." are
: mapped to files without any extension.
: 
: $ touch ab.
: $ ls
: ab
: $ rm ab.
: $ ls
: $ 
That's well-known DOS-like behaviour. Not the bug I described. See that 
the file created was named "ab" without a dot.

The problem is rather the following:
$ touch ab
$ ln ab xy.
$ ls
ab      xy.
$ rm ab
$ ls
xy.
$ rm xy.
rm: xy.: No such file or directory
$ rm xy
rm: xy: No such file or directory
$ ls
xy.
$ grrr


: It sounds like this is a "feature" of the Win32 file system, in that
: it doesn't *quite* preserve the file name correctly.  If so, there's
: not much we can do about it.  Except, perhaps, to MIME encode the
: problem file names :-(
The issue is not one of missing filename preservation either, rather on 
the contrary.

: DJ Delorie wrote:
: > It sounds like this is a "feature" of the Win32 file system, in that
: > it doesn't *quite* preserve the file name correctly.  If so, there's
: > not much we can do about it.  Except, perhaps, to MIME encode the
: > problem file names :-(
: It's a "feature" of the - in this respect absolutely brain damaged -
: NT/Win32 subsystem  file naming/handling: 
: NTFS knows the difference between "xy." and "xy" , but the Win32 layer 
: "erases" this knowledge !
: Getting this NTFS capability back under Win32 is not a trivial task,
: perhaps not doable without writing a kernel subsystem .
Don't know if that's the situation. I tried every tool I have (Norton 
commander clones etc.)

: > Getting this NTFS capability back under Win32 is not a trivial task,
: > perhaps not doable without writing a kernel subsystem .
: > P.S.: The Interix Posix subsystem knows the difference on NTFS
: We've talked about bypassing the win32 file system layer for other
: things, but decided against it because it lets you create files that
: other standard win32 programs (like explorer) can't deal with.
Which one can?

Regards,
Thomas Wolff

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]