[Nexus-developers] Hmmm... nxdiff bug...? (or operator error... :-)

[James Arthur Kohl]

Hey Gang!  :-)

I just got an error running "nxdiff" as part of a test suite:

   % nxdiff -L -S --show-values SEQ_3733_m1p.nxs ../GoldenData/SEQ_3733_m1p.nxs
   < /entry-bad/SNSHistoTool/SNSbanking_file_name
   < /entry-bad/SNSHistoTool/SNSmapping_file_name
   < /entry-bad/SNSHistoTool/author
   < /entry-bad/SNSHistoTool/command1
   < /entry-bad/SNSHistoTool/date
   < /entry-bad/SNSHistoTool/description
   < /entry-bad/SNSHistoTool/version
   < /entry-bad/duration
   < /entry-bad/end_time
   | /entry-bad/instrument/bank100/pixel_id ATTRIBUTES MISMATCH: {'target': '/entry-bad/instrument/bank100/pixel_id'} != {}
   Traceback (most recent call last):
     File "/SNS/software/bin/nxdiff", line 558, in <module>
       shownandiffs=options.shownandiffs)
     File "/SNS/software/bin/nxdiff", line 384, in printDataDiff
       if (left.type == right.type) and (left.dims == right.dims):
   ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()
   52.930u 3.189s 0:56.49 99.3% (0t+0ds+0av+0mx)k 0i+0o (0mj+364971mn)pf 0swp

The new file (on the left) has additional fields & attributes
not present in the former Golden Data, so --show-values is
apparently croaking when faced with comparison to an "empty"
(missing) array...  ;-}

Has anyone else seen this...?  Already fixed in next release...?  :-)

(Fingers crossed... ;-)

Please let me know.  Thx!  :)

	Jeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeem ;)

P.S. Just for yuks, I reversed the order of the files, but same-o same-o... :-)