Merge pull request #351 in CASA/casa6 from CAS-13334 to master
* commit '5402da0a44402041405ada6b2682f9c1baa87b92':
Forgot that old vi is still being tested in synthesisimager
fix a type that got introduced in the negative frequency case
Missed a compiler warning about a non return value
Fixed a first partition swing underestimation due to hitting channel 0 and some cleanup of temp files that seems to remain sometimes
fixed a couple of extra commenting
Fix an issue when using multiple ms and perchanweightdensity=True... the imaging weight number of channels was not accounting for the retuning of spw in the second and subsequent ms