assistant: Fix a race condition that could cause a pointer file to get ingested into...

assistant: Fix a race condition that could cause a pointer file to get ingested into the annex

This was caused by commit fb8ab2469d389e5b1e554831eeb8b7c7a072d5d7 putting
an isPointerFile check in the wrong place. So if the file was not a pointer
file at that point, but got replaced by one before the file got locked
down, the pointer file would be ingested into the annex.

The fix is simply to move the isPointerFile check to after safeToAdd locks
down the file. Now if the file changes to a pointer file after the
isPointerFile check, ingestion will see that it changed after lockdown,
and will refuse to add it to the annex.

Sponsored-by: the NIH-funded NICEMAN (ReproNim TR&D3) project