On Nov 7 15:41, Christian Franke wrote: > Corinna Vinschen wrote: > > On Nov 7 13:06, Christian Franke wrote: > > > According to > > > http://sourceware.org/ml/cygwin/2007-05/msg00359.html > > > http://sourceware.org/ml/cygwin-patches/2006-q4/msg00008.html > > > Cygwin includes support to mount a //?/GLOBALROOT/... path > > > > No, it doesn't. It allows to *access* the path, but the mount call > > follows a somewhat different route. You can only access the path > > directly (a concenssion to a necessity which has no POSIX equivalent) > > or, as you did, map the VSS volume to a drive letter. The patch is > > not in 1.5.25, btw., only in 1.7.0. > > > > Thanks for clarification. > > The access to the path works, except the access to the root dir of the > volume. It appears as a regular file which apparently provides raw read > access to the partition: > > $ ls -l '//?/GLOBALROOT/Device/HarddiskVolumeShadowCopy1' > -rw-r--r-- 1 ... 0 Dec 1 2006 //?/.../HarddiskVolumeShadowCopy1 > > $ xxd '//?/GLOBALROOT/Device/HarddiskVolumeShadowCopy1' > 0000000: eb52 904e 5446 5320 2020 2000 0208 0000 .R.NTFS .....
The windows path handling code in Cygwin doesn't recognize the path as the path to a drive's root dir. The path normalization code removes all trailing backslashes. When accessing a root dir, Windows requires a trailing backslash. When accessing the volume without the trailing backslash, it's opened for raw access. This is still a highly unusual situation from a POSIXy point of view. I'm not sure I want to slow down Cygwin for this any further. It seems the ability to mount a shadow volume as a dos drive is sufficient for this kind of problem, isn't it? Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
