confirmed. I saw zero fails and overlooked the 4 regressions on the make check for bug #66025.

The buildbots are failing with the following in the logs:

>>>>> processing /home/buildbotu/fc25-x86_64/gcc-lto-fedora/src/scripts/statistics/movstd.m
***** assert <*66021> (movstd (1:5, 3, []), [1/sqrt(2), 1, 1, 1, 1/sqrt(2)], eps)
!!!!! regression: https://octave.org/testfailure/?66021
movfun: DIM must be a positive integer-valued scalar
***** assert <*66021> (movstd (1:5, 3, [], 2), [1/sqrt(2), 1, 1, 1, 1/sqrt(2)], eps)
!!!!! regression: https://octave.org/testfailure/?66021
movfun: DIM must be a positive integer-valued scalar
>>>>> processing /home/buildbotu/fc25-x86_64/gcc-lto-fedora/src/scripts/statistics/movsum.m
>>>>> processing /home/buildbotu/fc25-x86_64/gcc-lto-fedora/src/scripts/statistics/movvar.m
***** assert <*66021> (movvar (1:5, 3, []), [0.5, 1, 1, 1, 0.5], eps)
!!!!! regression: https://octave.org/testfailure/?66021
movfun: DIM must be a positive integer-valued scalar
***** assert <*66021> (movvar (1:5, 3, [], 2), [0.5, 1, 1, 1, 0.5], eps)
!!!!! regression: https://octave.org/testfailure/?66021
movfun: DIM must be a positive integer-valued scalar

Could that be related to these changes?

while working on creating BISTs for the mov* functions, I noticed that movvar and movstd treated a [] for OPT as a 1 and not a 0 (the stated default). this is also inconsistent with var & std and incompatible with matlab. In addition there was no real input validation and other things (nonscalars, etc) were allowed to pass through opt as well.

finishing a patch for stable that adds input validation on opt, corrects the default assignment, and adds the FIXME requested general function BISTs to all mov* functions.