More Unicode 16 linebreaking changes

[eggrobin]

I had missed this one in #798 because we were missing a LineBreakTest.txt action item for it:

[179-C25] Consensus: In UAX #14, change rule LB21a from HL (HY | BA) × to HL (HY | BA) × [^HL] (do not break after the hyphen in Hebrew-hyphen-non-Hebrew) to match the rationale for adding LB21a given in L2/11-141R. For Unicode Version 16.0. See document L2/24-064 item 5.2.

[eggrobin]

Note: running the ICU rbbi/RBBITest/TestUnicodeFiles on the current unicodetools LineBreakTest.txt (patched to replace U+1B50 with U+1BC0 since ICU does not yet have its LB change UTC-177-C46) on my uax14-integration branch failed with

         Test file "LineBreakTest.txt", line 16664, unexpected break found at position 2
         Expected : ×א×-÷א÷
         Actual   : ×א÷-÷א÷

So our conformance tests would have caught that blunder during ICU integration.

With the LineBreakTest.txt from this branch (similarly patched) the ICU test now passes.

Note that this all passes after removing the exemptions for ICU-21097, so this one can be closed soon (with the resolution that Andy had predicted four years ago).