Ran out of edit time, gang. Corrected version follows.
----
To elaborate on parse-time vs run-time errors in qp2,
----------------------
b := (S*Sxy - Sx*Sy)/(S*Sxx - Sx*Sx);
if b = 0 then print symbol,"b was zero with j: ",j; endif;
pctgain := abs((b1/b) -1);
--------------------
fails with a parse-time div by 0 error, even though commenting out the pctgain line shows b isn't 0 at run time.
The parse-time error is fixed by
-----------
b := .0000001+(S*Sxy - Sx*Sy)/(S*Sxx - Sx*Sx); // kludge
-----------
or by enclosing pctgain in if ... endif.
-------------
b := (S*Sxy - Sx*Sy)/(S*Sxx - Sx*Sx);
if b != 0 then
pctgain := abs((b1/b) -1);
endif;
-------------
It is also amusing that, though numdays is rejected at parse-time in Jan's
------
NumDays:= -(daysloaded -1);
println NumDays, Max(0, NumDays, hi);
-------
negating NumDays by
-------------
println NumDays, Max(0, -NumDays, hi);
-----------------
works fine, but it always returns zero.
That is, +255 is accepted by Max, but it punts with zero since it can't predict the future ;)
Can you easily fix these parse-time check errors, Gary?
Bob
|
