Corrected cta-admin af ls command so that it fills the superseded vid and superseded columns in the result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Added a check while mounting a tape for writing : if some tape files located in the writing tape are not superseded by others,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.

cta-admin was failing due to incoherent internal data.

This included extracting missing data while popping elements from archive queue.

Conflicts:
	scheduler/SchedulerTest.cpp

- Queueing into the RetrieveQueueToReportToRepackForFailure
- Reporting of the failed RetrieveJobs (add statistics + deletion of the failed sub requests)
Unit test associated : expandRepackRequestFailedRetrieve

These requests are queued in the RetrieveQueueToReportToRepackForFailure
Adapted the Sorter to be able to queue a RetrieveRequest from Repack into a queue identified by the RepackRequest's address

Someone added xrootd in eos-citrine-depend repo which had a priority of 2. Lowering this one as people are messing with it.

repositories.

basically requiring xrootd-4.8-3 and not our 4.9.0 this is an externally
triggered problem. To fix this we can either disable protectbase yum
plugin for all our yum commands, remove the plugin, sed -i
's/protect=1/protect=0/' /etc/yum.repos.d/* in all our containers OR
enable baseprotect on our repos...
I enabled baseprotect on all our repos.

Updated the unit tests to validate the queues and requests are as expected.

Changed the reporting structures in the repack request to better handle repoting
(we will report one archive job per rearchived copy number, which is potentially
several per fSeq.

Created the getNextBatch() function for retrieve successes and partially implemented
the report function.