Fix side effect in the affine method of relabelled cartan types

[nthiery]

This fixes the following side effect in the affine method of relabelled cartan types::

sage: C = CartanType(["A",3]).relabel({1:2,2:3,3:1})
sage: C
['A', 3] relabelled by {1: 2, 2: 3, 3: 1}
sage: C.affine()
['A', 3, 1] relabelled by {0: 0, 1: 2, 2: 3, 3: 1}
sage: C
['A', 3] relabelled by {0: 0, 1: 2, 2: 3, 3: 1}

The patch also removes a couple trailing white spaces.

Depends on #4327
Depends on #13724

CC: @tscrim @anneschilling

Component: combinatorics

Author: Nicolas M. Thiéry

Reviewer: Anne Schilling

Merged: sage-5.10.beta4

Issue created by migration from https://trac.sagemath.org/ticket/14576

[nthiery]

comment:1

Anne: that was the cause of the failure for the non symmetric macdonald polynomials in type G2.

[nthiery]

comment:2

All long tests passed on my machine.

[anneschilling]

comment:3

Hi Nicolas,

Great! Could you please remove one semicolon after TESTS:: . Other than that it indeed looks good!

[nthiery]

comment:4

Thanks for catching this!

Done and pushed. Hence positive review!

[nthiery]

Reviewer: Anne Schilling

[jdemeyer]

comment:6

attachment: trac_14576-cartan_type-relabel_affine-nt.patch is missing a username in the patch metadata. Did you create the patch using Mercurial's hg export?

[nthiery]

Attachment: trac_14576-cartan_type-relabel_affine-nt.patch.gz

[nthiery]

comment:7

Replying to @jdemeyer:

attachment: trac_14576-cartan_type-relabel_affine-nt.patch is missing a username in the patch metadata. Did you create the patch using Mercurial's hg export?

Oops, sorry. I guess I did not! Kind of weird: given my local shortcuts it's actually harder for me not to use hg export. Whatever, this is fixed now.

Thanks for catching this!

[jdemeyer]

Changed dependencies from #13724 to #4327, #13724

[jdemeyer]

Merged: sage-5.10.beta4