Warnings in plot_slope_field

[novoselt]

Duplicate of #11208

If I do (in Sage 4.7.2 or 4.8.alpha6)

plot_slope_field(1, (1,2), (3,4))

(or anything else that I have tried), then in front of the plot, which looks as expected, I get

Warning: divide by zero encountered in divide
Warning: invalid value encountered in multiply
Warning: invalid value encountered in multiply
Warning: divide by zero encountered in divide
Warning: invalid value encountered in multiply
Warning: invalid value encountered in multiply

My investigation so far shows: this function is a little wrapper for plot_vector_field which sets new defaults for three options. If headlength is set to 0, there are warnings.

A possible workaround is to set it to something small: 0.001 seems to work on the screen. However, since with headaxislength=0 and headlength=0.2 I definitely see "a hair on the end", it does not seem to be satisfactory from a perfectionist point of view (and producing vector formats that may be scaled or printed with high resolution).

Besides, current settings do produce the accurate plot, it just shows warnings that should not be there, and I have no idea how to fix it.

CC: @jasongrout

Component: graphics

Reviewer: Karl-Dieter Crisman, Andrey Novoseltsev

Issue created by migration from https://trac.sagemath.org/ticket/12277

[jasongrout]

comment:1

It seems like I remember that this was fixed in the new matplotlib 1.1.0 release (an spkg is up on trac somewhere).

[jasongrout]

comment:2

Ah yes, #11915. Do you still see the problem after applying that spkg?

[novoselt]

comment:3

Replying to @jasongrout:

Ah yes, #11915. Do you still see the problem after applying that spkg?

Yes, the same issue both in 4.7.2 and 4.8.alpha6 with #11915 applied. Are these warnings coming from matplotlib? Also, what would be an appropriate doctest for this bug?

[kcrisman]

comment:4

This is a dup of #11208? Probably. See also #2922.

[novoselt]

comment:5

Yes, strange I couldn't find it, I tried to search... My current workaround is that I have redefined plot_slope_field. It is in a way good, since I also had to fix aspect ratio (for 4.7.2) and changed defaults for colour and frame to match the textbook I am using.

[novoselt]

comment:7

Am I supposed to do anything with author/reviewer fields?

[kcrisman]

comment:8

One puts the reviewers only to those who realized, it, at least in the usage lately. If I misspelled your name, just fix it.

[kcrisman]

Reviewer: Karl-Dieter Crisman, Andrey Novoseltsev