Fixed handling of exception when starting new job : we should not interpret them as coming from the transfermanager

for operational convenience.

updated release notes, taking into account the postupgrade script and informing about the need for proper migrationRetry policy

Improved error reporting for early errors in the stagerJob. See bug #90086: lack of error reporting when stagerJob fails at the very beginning

bug #91761: RFE: default migration retry policy should not retry when the file is gone from the name server

Changing the default migration policy to no retry migartions for file not in name server.

plus fixed some typos.

is configured to have mounts with a small minimal amount of data,
it could happen that a single large file triggers the creation of
many mounts. This is now prevented.

Also, partial refactoring of nsrm client to improve readability

Fixed a stack-overflow bug in the CASTOR 2.1.12-2 internal release of the
tapebridged daemon.  The bug was caused by an unintentionally recursive
function.  The bug was not detected on my development system but was detected
on castorcert5 because the bug only manifested itself when an error was
reported to the tapebridged daemon by the rtcpd daemon.

This commit is the first version of the tapebridged daemon that requests more
than one file for migration or recall from the tapegatewayd daemon.

Readded the chmod test that was temporarily removed in commit 23538 while doing the release 2.1.12-1