[with patch, positive review] Improve doctest coverage for sage/modular

[davidloeffler]

The attached patch adds doctests for 28 previously undoctested functions in the sage/modular directory, and fixes 2 small bugs uncovered in the process: one in pickling of arithmetic subgroups defined by permutations, and one in dirichlet characters (galois_orbits() returned meaningless garbage when the base ring wasn't an integral domain).

This brings the doctest coverage to 100% for everything *except* the three big subdirectories modform/, modsym/ and hecke/. I will get to work on these next.

[davidloeffler]

patch against 3.4.1.rc1

[mabshoff]

Let's change the status so the right reports pick up this ticket :)

Cheers,

Michael

[was]

REVIEW:

• Put backquotes aroudn start_weight in the modform_generators docstring: - start_weight -- an integer (default: 2)
• A doctest fails on 32-bit OS X:

  sage -t --long devel/sage/sage/modular/arithgroup/arithgroup_perm.py
  **********************************************************************
  File "/Users/wstein/build/sage-3.4.1.rc1/devel/sage-main/sage/modular/arithgroup/arithgroup_perm.py", line 202:
      sage: cmp(G, 1)
  Expected:
      -1
  Got:
      1
  **********************************************************************
  1 items had failures:
     1 of   6 in __main__.example_9
  ***Test Failed*** 1 failures.
  

I recommend changing the doctest to:

   sage: cmp(G,1) in [-1,1]

since it depends on the OS.

These are trivial changes, so I've posted a tiny patch that adds them and given this a positive review.

[was]

apply this after applying the above patch

[mabshoff]

Merged both patches in Sage 3.4.1.rc2.

Cheers,

Michael

[davidloeffler]

Here's some more -- mostly in sage/modular/hecke/hecke_operator.py and sage/modular/hecke/module.py. This patch also adds Brandt modules into the reference manual.

[mabshoff]

Please do not reopen tickets with merged patches. Instead open a new ticket for the new patch. I have deleted the new patch.

Cheers,

Michael