Unnecessary coordinate check while initializing an integrated curve

[egourgoulhon]

In Sage 8.9, we have:

sage: E.<r,phi> = EuclideanSpace(coordinates='polar')
sage: p = E((1, 0))  # the initial point
sage: v = E.tangent_space(p)((2, 1))  # the initial vector
sage: t = var('t')
sage: c = E.integrated_geodesic(E.metric(), (t, 0, 10), v)
...
ValueError: initial point should be in the domain of the chart

Now, the initial point p does lie in the domain of the chart:

sage: E.default_chart()(p)
(1, 0)

Indeed, 0 is a perfectly valid value for the polar coordinate phi:

sage: E.default_chart().coord_range()
r: (0, +oo); phi: [0, 2*pi] (periodic)

The above error turns out to result from a too strict and unnecessary coordinate check performed in IntegratedCurve.__init__(). This check is removed here.

CC: @sagetrac-karimvanaelst

Component: geometry

Keywords: integrated_curve, geodesic

Author: Eric Gourgoulhon

Branch/Commit: c0efb2e

Reviewer: Karim Van Aelst

Issue created by migration from https://trac.sagemath.org/ticket/28669

[egourgoulhon]

Commit: 3ab55d7

[egourgoulhon]

Branch: public/manifolds/integrated_curve_init

[egourgoulhon]

New commits:

3ab55d7

Remove unnecessary coordinate check in IntegratedCurve.__init__()

[fchapoton]

comment:2

please add a doctest

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

c0efb2e

Add doctest in IntegratedCurve.__init__()

[sagetrac-git]

Changed commit from 3ab55d7 to c0efb2e

[egourgoulhon]

comment:4

Replying to @fchapoton:

please add a doctest

Done in the above commit.

[sagetrac-karimvanaelst]

comment:5

Replying to @egourgoulhon:

Replying to @fchapoton:

please add a doctest

Done in the above commit.

Ticket seems ready to be closed.