Rename logptdx to sinint-logptdx and return it from dprog
An examination of the code shows that `dprog` sets `logptdx` so we're
assuming its the main routine for initializing `logptdx`. So make
`dprog` return two values: the original return value and a second
value of `logptdx`.
Also `fprog` appears to initialize `logptdx` and calls `dprog`, so
modify `fprog` to handle the two return values from `dprog`
appropriately.
Then rename `logptdx` to `sinint-logptdx` to make it distinct from the
specvar `logptdx`.
Finally, in risch.lisp, `rischfprog` calls `dprog` so update the code
to handle the second return value of `dprog` and set it to
`logptdx` (which we didn't rename in risch.lisp). That's only place
outside if sinint.lisp to call `dprog`.
More minor reindentation make some things easier to read
No functional changes.
a1ee2b0...
by
Robert Dodier <email address hidden>
In save, evaluate filename argument outside of WITH-MAXIMA-IO-SYNTAX.
Fixes SF bug #4097: "Bad string representation of integers in the first argument of `save`"
ccf37eb...
by
Robert Dodier <email address hidden>
Increase right-binding power of MMINUS from 100 (same as MPLUS) to 134,
which is greater than left-binding power of MTIMES (namely 120).
This discourages unnecessary parentheses around the leading term of a "+"
expression when that term has a minus sign.
Possible fix for SF bug #4090: "output formatting: -x-1 has redundant parentheses: (-x)-1"
MMINUS appears in expressions prepared for output by NFORMAT.
The binding power for $- which is recognized by the parser, is unchanged.
Therefore this commit only changes the display of minus expressions,
not input.
There is still some residual strangeness in formatting expressions with minus signs,
which stems from NFORMAT reorganizing (- x)*y to -(x*y).
Printing parentheses for the reorganized expression is actually correct,
because the parser yields (- x)*y, not -(x*y), for the input -x*y.
I believe at this point the focus should shift to NFORMAT.
Rename exp in ratint and remove special decl for exp
By renaming the `exp` arg of `ratint` to `sinint-exp`, we have removed
the last use of the special var `exp` from sinint.lisp. Thus, we can
also remove the special decl of `exp`.