vpath search failures in Clearcase when files do actually exist

[Reini Urban]

Pierre A. Humblet schrieb:
> On Wed, Nov 10, 2004 at 01:09:47PM -0500, Christopher Faylor wrote:
>>On Wed, Nov 10, 2004 at 12:47:26PM -0500, Pierre A. Humblet wrote:
>>>On Tue, Nov 09, 2004 at 07:51:23PM -0500, Christopher Faylor wrote:
>>>>On Tue, Nov 09, 2004 at 07:21:00PM -0500, Jeremy Broughton wrote:
>>>>
>>>>>Cyg ver:    1.3.10 (although I have reproduced the same problem with 1.5.7)
>>>>>OS : Windows 2000
>>>>>File system:   MVFS (running gnumake inside a Rational Clearcase dynamic
>>>>>view)
>>>>>cygcheck:   (See attached file: cygcheck.out)
>>>>>
>>>>>It appears as though there is a problem using the /cygdrive/m type paths
>>>>>inside MVFS when there are extremely large dependency trees.
>>>>
>>>>Unless you can point to problems with a standard, supported W2K
>>>>filesystem, please don't expect anyone to try to fix this.
>>>>
>>>>Cygwin uses standard Win32 APIs for detemining if something is a
>>>>directory.  If those APIs are not working for Clearcase then it is
>>>>unlikely that anyone will be motivated to fix the problem on the Cygwin
>>>>side.
>>>
>>>The OP was suggesting both a problem determining if something is a
>>>directory and a problem related to pathnames.  Running strace my shed
>>>some light on the root cause.
>>
>>Would it have made you feel better if I had said "for dealing with paths
>>and determining if something is a directory"?
> 
> 
> The "related to pathnames" was refering to the last paragraph in the initial
> post, which points to Cygwin (although I can't imagine how):
> " As shown above, using vpath .  ../..  /cygdrive/m/top fails, even
> though it should work.  Using vpath .  /cygdrive/m/top fails with the same
> errors as above.  However, vpath .  M:/top succeeds in the dependency
> check, even though M:/top should apparently be written as /cygdrive/m/top
> to comply to cygwin standards."
>   
> 
>>If the OP is willing to do this then I hope he does it on a standard
>>filesystem like NTFS, as I originally suggested.  If there is a problem
>>there then it is obviously not a ClearCase problem.
> 
> 
> Right. But then if it is a ClearCase problem, it can't be identified...
> My request would be to trace under the latest Cygwin.

BTW: Can someone provide testcode to identify if the current drive as 
MVFS please. At least I think that a clearcase view is mounted as 
driveletter.
Does it show up as GetLogicalDrives()? (should so)
In detail, what does GetVolumeInformation().lpFileSystemNameBuffer 
return for such a MVFS drive?

For the proposed mounttable (getmntent) changes.
   http://cygwin.com/ml/cygwin/2004-10/msg01449.html
     user|system => mnt_opts
     GetVolumeInformation().lpFileSystemNameBuffer as mnt_type.
-- 
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/

--
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/