Forgot to add the error strings for the following two error codes (thanks Dennis): SENOPORTINRANGE and SENOVALUE

Run cleanup in stageForcedRm even when no diskcopy is invalidated, as subrequests and tapecopies may still be cleaned up. Fixes Bug #54480

different file. This code is not meant (yet) for a production release
of the Castor nameserver.

a daemon may lead to bogus insertions in the list of thread pools,
which then provoke segmentation faults once the threads are started.

The fix is 'tentative' because I need to confirm it and its consequences with Dennis.

Fixed the bug where writetp would correclt detect a bas POSITION parameter, but would incorrectly print the actual command-lin argument given.  The bug was a copy-paste bug which resulted in the PSOTION parameter's value being treated as the value of an optional command-line argument, i.e. optarg, when it is now in fact a non-optional argument, i.e. argv[optind]

In case of readtp if filepath do not identify a directory, throw an exception.
In case of writetp if filename identify a directory, throw and exception.

Fixed a bug introduced on "Tue Aug 11 19:29:36 2009" when code was added so that "umask is now passed from readtp through to rtcpd".  The problem was a copy-paste bug.  I used the same code for sending the first file to migrate to RTCPD and for sending subsequent files.  The first file should be sent using the initial RTCPD callback connection (which the code was doing), but for subsequent files the Disk IO callback connections should have been used (but the code was always using the initial RTCPD callback connection).  writetp tests were passing because we were only ever writing one file.  The tape gateway tests of Giulia were correctly failing because her tests tried to write more than one file

DLF_LVL_ERROR and DLF_LVL_DEBUG.

writetp now checks that the user is the owner of the tape in question's pool or that they have ADMIN privileges

tape copies of a file in case of a successful recall where missing tape copies were detected

readtp now RFIO stats the first destination directory and writetp now RFIO stats the first source file.

The tpcp commands will now try to delete their VDQM volume requests if any error is detected.  Of course this will not always succeed, beacuse either the request is no longer in the system or it is assigned to a drive

bug #54252: VDQM does not allow a user to delete their own request