I just got through my first successful build+test on the Mac in a while and 
notice the ‘new-to-me’ regress-dem hawaii test is failing.  Looking into it, 
I’m not seeing any way to debug/diagnose what’s wrong because it’s only got an 
expected md5sum:

dem results differ for 
/Users/morrison/brlcad.trunk/.build/regress/gcv/dem/hawaii-c.dsp: expected 
80f98f409514586e0be12884e2118230, got 064766d19922b1e441ca5e069dfca10a

Anyone else seeing this?  I don’t know if it’s a false positive or not, but 
there is a DSP of hawaii in the file and it looks fine.

That leads me to two thoughts, first being that the test is too sensitive as it 
will fail on essentially any db layer change unrelated to gcv/dem conversion.  
I suggest some sort of analytic, round-trip, or render test instead.  Second 
thought is more obvious, that it would also be good to know exactly why the 
output is different even if the test is improved, but I need the expected .g to 
investigate bytes.

Cheers!
Sean



_______________________________________________
BRL-CAD Developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to