I previously had not looked at the further discussion in that report,
but the discussion of branches of the log function and the Sage
"simplification" shows it is not a parsing problem.
Since log() is in general multivalued, any answer that
ignores this possibility may fall into a trap. Setting a
I just tried the problem in Maxima 5.40.0 ; the result is correct as I
assume
Emmanuel also discovered. The "fix" suggested by the poster of Trac#28431
(to divert some class of integral problems to sympy) does not strike me
as plausible.
Maybe something about Sage parsing of a minus sign??
RJF