-
A prior #52 (closed) commit ebea4808 disabled the
tnodata3
subtest because it started failing due to the code changes in that commit. Instead the subtest should have been adjusted to verify the new code behavior. This comment does that. -
While adjusting the subtest behavior, I found the comments in the subtest did not help understand what was being tested. The comments described nodes of the form
^ORD(100.01,14,...)
whereas the data verification that happened later in the test used asserts with@aimgbl@(1,"",1)
and was not verifying@aimgbl@(1,"",14)
. Also, it was not clear where the data being described in the comments were loaded. -
Therefore I simplified this test to load the data right here using some
^tnodata3
nodes. -
While at that, also enhanced the test to use 2 levels of fixed subscripts
"const1"
and"const2"
. And ensured different subscripts before the fixed subscripts exercised different combinations where the first and or second fixed subscript existed or not. -
And enhanced the test to verify not just what we expect to be seen in the xref but also what we do not expect to see in the xref (using a
1=$data(@aimgbl@(0))
check at the end). -
Also verified this reworked
tnodata3
subtest does fail (as expected) with the following error if I revert the prior #52 (closed) commit ebea4808.tnodata3^%YDBAIMTEST - indexing fixed nodes that don't exist - Error from tnodata3+25^%YDBAIMTEST