Removed helgrind error by removing a linkage in the tapeserver/castor/tape/tapeserver/file/CMakeLists.txt file

Refactored the async deletion of repack subrequests files via XRootD (used the xroot's async deletion)
Deleted the folder containing the repack subrequests files (Local file + XRootD)

The folder containing the files need now to be removed

Added the creation of the directory corresponding to the vid of the tape to repack (xroot + local file system)

workaround for jq when output is not a tty as seen [here](https://github.com/stedolan/jq/issues/1110#issuecomment-193504718)

Added a display of the state of the tape before and after each repack

Simple repack on CI

Corrected cta-admin af ls command so that it fills the superseded vid and superseded columns in the result

Replaced the map of tape files with a list as the assumption that we have one copy per copyNb is no more valid with repack.

Added a check while mounting a tape for writing : if some tape files located in the writing tape are not superseded by others,
the write session will fail.

Note: The test CatalogueTest.exist_non_superseded_files_after_fseq is supposed to fail.

Upgrading to eos 4.4.32 that brings some [SSD scrubbing fixes](https://its.cern.ch/jira/browse/EOS-3349) for the new pps and xrdfs query backup bit on eos.

cta-admin was failing due to incoherent internal data.