bug somewhere in the symbolics

[williamstein]

This is from Hector:

I also hit this bug while doing this (taken from the "piecewise"
documentation):

sage: f1 = lambda x:-1
sage: f2 = lambda x:2
sage: f = Piecewise([[(0,pi/2),f1],[(pi/2,pi),f2]])
sage: P = f.plot_fourier_series_partial_sum(15,pi,-5,5)   # long time
boom
...

/Users/was/s/local/lib/python2.5/site-packages/sage/calculus/calculus.py in <lambda>(i)
   3607             # We need to do this maximum to correctly handle the case where
   3608             # self is something like (sin+1)
-> 3609             n = max( max(map(lambda i: i.number_of_arguments(), self._operands)+[0]), len(variables) )
   3610         self.__number_of_args = n
   3611         return n

<type 'exceptions.AttributeError'>: 'Pi' object has no attribute 'number_of_arguments'

Component: calculus

Issue created by migration from https://trac.sagemath.org/ticket/1772

[mwhansen]

comment:2

Attachment: 1772.patch.gz

[williamstein]

comment:3

I also fixed this in the same way earlier today (as part of one my other patches), but my patch was just a few lines to actually fix the listed problem. The patch attached to this ticket, fixes the problem and does a HUGE amount more to vastly improve doctesting in some files, etc. I.e., this is great.

I have not fully reviewed the patch yet, though I've looked it over by eye and it looks very good.

[rlmill]

Attachment: 1772-bug.patch.gz

Fixes the one doctest failure in constant.py

[sagetrac-mabshoff]

comment:5

Merged in Sage 2.10.1.alpha0