Show the error messages from errors encountered when plotting
Which error messages get printed when the plot functions catch an error
are the same as which error messages get printed when errcatch catches
an error. The printing is controlled by errormsg.
This was discussed in bug report #3881.
See also commit a7912deb.
No problems with the test suite or share test suite.
Fix bug #3881: plot2d not ignoring errors within functions
The functions constructed by COERCE-MAXIMA-FUNCTION-OR-MAXIMA-LAMBDA
and %COERCE-FLOAT-FUN have been inconsistent wrt error-catching. The
bug shown in the bug report is due the function constructed by
COERCE-MAXIMA-FUNCTION-OR-MAXIMA-LAMBDA not catching general Maxima
errors (signaled via MERROR), which caused no plot to be generated.
Now both of these functions use the ERRCATCH macro to catch errors.
With this, the constructed functions are consistent and they catch
any errors that errcatch would catch.
No problems with the test suite or share test suite. New tests have
been added to rtest_plot.
We get a lisp error when evaluating li[3](-.862) because we need to
compute li[3](1), but li3numer doesn't know how to convert 1 to a
float. Adjust the call (li3numer 1) to be (li3numer (float 1 x))
because at that point, we have the float value x so we can convert 1
to the appropriate float type.
Fix li[3] returning complex values when it should be real.
Tentative fix. Get rid of the case using the identity using 3 li[3]
terms. Use the series using powers of logs instead to compute the
result. Then take the real part of the result if the arg is real
because the result should be real.
