Stagnant regression test output

[Dan Eble]

Last night, I noticed that the log output for my local regression tests did not 
hold the updated output I expected after test-clean, check.  It still held old 
output that did not match what my rebuilt lilypond binary produced in a manual 
run on the test input.  I ended up deleting my whole build directory and 
starting over.

At this point, I don’t have any reason to believe that there is a problem 
affecting other people; however, I don’t have evidence to the contrary either.  
It’s going to take me some time to get to the bottom of it.  My hypothesis is 
that there’s yet another undiscovered problem with my using a separate, 
read-only source directory, but I want to warn everyone to pay attention as you 
test, just in case it’s not limited to my environment.

Regards,
— 
Dan